MXPA99008689A - Image transmitting method, image processing method, image processor, data storage medium - Google Patents

Abstract

An image transmitting method, wherein compressed image data (Dv) obtained by compression-coding digital image data corresponding to one motion picture and including an identification flag (Hfd) representing whether or not the compressed image data (Dv) is appropriate for random independent reproduction of an arbitrary image and sent next to the synchronizing signal (Hsd) at the first portion of the header (Hv). On the reproducing side where the compressed image data transmitted by such an image transmitting method is reproduced, during the analysis of the header (Hv) given to the compressed image data (Dv) corresponding to one motion picture, the appropriateness of the random independent reproduction of the compressed image data (Dv) is sensed in a short time by analyzing the identification flag (Hfd).

Description

METHOD OF IMAGE TRANSMISSION, METHOD AND APPARATUS OF PROCESSING OF IMAGE AND MEANS OF DATA STORAGE TECHNICAL FIELD The present invention relates to an image transmission method, an image processing method, an image processing apparatus with a data storage medium, and more particularly to a method for transmitting coded digital image data corresponding to an image comprising a plurality of frames, a method and an apparatus for encoding the digital image data, a method and an apparatus for decoding the encoded digital image data, and a data storage medium which contains a program for implement coding and decoding processes for digital image data by using a computer.

ANTECEDENTS OF THE TECHNIQUE In order to efficiently store or transmit digital image information, ie, image data of a digital signal, it is required that the digital image information be compressively encoded. As the available methods REF. 31256 to comprehensively encode the digital image information, there are the waveform coding methods such as subband, fractal wave train, and so on, as well as DCT (discrete cosine transform) typical of an image processing technique according to with JPEG (in English board of experts group of photographic coding) or MPEG (in English group of experts in films). Meanwhile, one method for removing redundant image information between adjacent and similar frames is to perform prediction between frames using motion compensation by representing pixel values in a current frame by difference between these pixel values and the pixel values of the pixels in a previous frame (past), and perform coding of a signal difference that corresponds to the difference. Next, an image coding method and an image decoding method according to the MPEG standard which performs a DCT process including motion compensation will be briefly described. In this image coding method, an input image signal is divided into a plurality of image signals corresponding respectively to a plurality of blocks (macroblocks) in a frame, and then the image signals are coded for each macroblock. A macroblock corresponds to a region of image display constituted of (16 x 16) pixels. When the input image signal corresponds to an object image, the image signal is divided into a plurality of blocks (macroblocks) that make up an exhibition region (object region) that corresponds to the object image in a frame. . The image signal corresponding to each macroblock is divided into image signals corresponding respectively to sub-blocks corresponding to the image display regions, each consisting of (8 x 8) pixels, and then the image signals are subjected to DCT process for each sub-block in order to generate DCT coefficients. Then the DCT coefficients are quantized to generate the quantization coefficients for each sub-block. The method for encoding the image signal corresponding to the subblocks in this way by the DCT process and the quantization process is called "coding scheme within the frame". At the receiving end, the quantization coefficients are quantized inversely and then subjected to an inverse DCT process for each sub-block in order to reproduce an image signal corresponding to the macroblock. The encoded data corresponding to a frame (image I) in which the image signal has been encoded by the coding method within the frame, can be reproduced independently. That is, they can be decoded without reference to the image data of another frame.

On the one hand, there is a coding method called frame coding scheme. "In this coding method, initially, a method is used to detect the movement of an illustration in a frame such as" block matching "to detect a region. constituted of (16 x 16) pixels with the least amount of errors between pixel values thereof and pixel values of a target macroblock that is to be encoded as a prediction macroblock, from an image signal corresponding to a coded frame which is temporarily adjacent to a frame to be encoded.Subsequently, the image signal of the prediction macroblock is subtracted from the image signal of the target macroblock to produce a signal difference of the target macroblock, which is It divides into a difference of signals corresponding to the sub-blocks, each consisting of (8 x 8) pixels. Signals are subjected to the DCT process to generate the DCT coefficients for each sub-block, which are quantized for each sub-block to generate quantization coefficients. The image signal corresponding to the object image is encoded between frames in a similar manner. At the receiving end, the quantization coefficients (quantified DCT coefficients) are quantified inversely and then subjected to the inverse DCT process for each sub-block in order to restore the signal difference of the macroblock. Then, from an image signal of a decoded frame, a prediction signal of an image signal corresponding to the target macroblock to be decoded is produced by motion compensation. Subsequently, the prediction signal and the restored difference signal are added to reproduce the image signal of the target macroblock. The encoded data corresponding to the frame (image P or image B) in which the image signal has been encoded by the frame coding method can not be reproduced independently. That is, it can not be decoded without reference to the image signal of another frame in the reproduction process. Subsequently, a structure of compressed image data (stream of bits) corresponding to a moving image consisting of a plurality of frames (illustrations) will be described. Figure 10 (a) shows an image data structure (moving illustration data) that corresponds to a moving illustration. A moving illustration comprises a plurality of frames. In Fig. 10 (a), the data D of the moving illustration comprises frame data P (l) -P (n) (n: natural number) corresponding to the respective frames.

Figure 10 (b) shows the structure of compressed image data Da within the frame obtained by performing the coding process within the frame to the respective frame data P (l) -P (n) which constitute the data D of illustration in movement. The compressed image data Da within the frame comprises Pa (l) -Pa (n) encoded frame data of respective frames and a Ha header comprising common data for these frames. Frames are I-coded illustrations within the frame. According to MPEG4, the Ha header is called a "VOL (video object layer)". Figure 10 (c) shows a structure of the compressed image data Db within frame that is obtained by performing the coding process within the frame for specified frame data of the frame data P (l) -P (n) and when performing the coding process between frames for the other frame data. The coding process between frames includes two types of processing. One is a predictive coding process which encodes an objective framework that will be encoded with reference to a previous framework (forward), and the other is a bi-directional predictive coding process which performs the coding of the objective frame when referring to previous and subsequent frames (forward and backward).

The compressed image data Db within the frame compressed frame data Pb (l) -Pb (n) of respective frames and an Hb header comprising data common to this data. As illustrated, the first frame of the illustration in motion is the illustration I encoded within the frame and the other frames are illustrations P which have undergone the forward predictive coding process or illustrations B which have undergone processes of bidirectionally predictive coding. Since the data Da of compressed image within the frame is produced by performing the coding process within the frame for each frame of the moving illustration without reference to another frame, it is suitable for use in random reproduction (decoding), although its efficiency Coding is relatively low. In other words, an advantage of the use of compressed image data within the frame is that the frames to be decoded are randomly selected and decoded immediately to reproduce an image. Particularly when compressed image data is edited, the image data compressed within the frame is easier to handle than the compressed image data between the frames. This is because the compressed image data within the frame occurs independently of other frame data but the compressed image data between frames does not.

On the other hand, since the compressed image Db data between frames is produced by performing the coding process between frames to almost all the frames of the moving illustration with reference to another frame and therefore its coding efficiency is high, they are less suitable for use in random reproduction (decoding). In the compressed image data Db between frames, when the decoding starts from the illustration P or the illustration B as the target frame to be decoded, it is necessary to decode a decodable frame independently present before the target frame. This is because the objective frame that is to be decoded is the frame which has been encoded with reference to another frame. For example, in the compressed image data within the frame, the frame data encoded Pae (1) -Pae (m) (m: natural number) which corresponds to 30 second data placed at the bottom of a one-hour moving illustration can be reproduced from the Pae (l) coded frame data at the start of this frame data (see Figure 10 (b)). On the other hand, in the compressed image data Db between frames, when encoded frame data Pbe (1) -Pbe (m) are reproduced corresponding to 30 seconds data placed on the back of the moving illustration of a time, the encoded frame data Pbe (l) at the beginning of this data can not be reproduced first (see Figure 10 (c)). The encoded frame data Pbe (l) can not be reproduced until the independently reproducible data (encoded frame data Pb (l) corresponding to the first frame of the moving illustration) through the encoded frame data present just before of the Pbe (l) data have been decoded. This is because the encoded frame data Pbe (l) is the data which has been encoded with reference to another frame. Meanwhile, a fast forward playback process can be performed while skipping S frames (S: natural number) to compressed image data within frames (see Figure 11 (a)). This is because the encoded frame data Pa (l), Pas (1) -Pas (f) (f: natural number) that are to be decoded in the fast forward reproduction process correspond to the I illustrations encoded within of frames which can be played independently without reference to other frame data. A fast rewind playback process as the inverse of the fast forward playback process can also be performed to the compressed image data within the frames in the same way.

On the other hand, in practice, the fast forward playback process can not be performed for the compressed image Db data between the frames (see Figure 11 (b)). This is because each of the frame data encoded Pbs (1) -Pbs (f) to be decoded in the fast forward reproduction process corresponds to the P picture encoded between frames of the frame-encoded B picture. . The data of respective encoded frames Pbs (l), Pbs (2), Pbs (3), ..., Pbs (f) can not be decoded until the corresponding waiting times tbl, tb2, tb3, ... tbf , i.e., times necessary to decode all of the encoded frame data present before the respective data Pbs (1) -Pbs (f) has elapsed. In other words, the encoded frame data Pbs (1) -Pbs (f) to be decoded in the fast forward playback process are played in the same synchronization as when they are played in a normal playback process. Accordingly, if a fast forward playback process is to be performed for the compressed image Db data between frames, the reproduced image resulting from the moving illustration will reproduce still images of the encoded frame data Pbs (1) -Pbs ( f) which are displayed sequentially at regular time intervals.

The fast rewind reproduction process can not be carried out with the image Db data compressed between frames, since the frame data encoded from the last frame can not be reproduced until all the frame data encoded has been decoded Each of the Ha and Hb headers of the data Corresponding compressed image Da and Db contain an identification flag which indicates whether the corresponding compressed image data is suitable or not for use in the independent reduction. As solutions to the problems associated with the compensation between the efficiency in compressively coding the image data and the suitability of the fast forward reproduction process, the following is conceived. The first solution is, as shown in Fig. 12, to store the compressed image data within frames suitable for use in the fast forward playback process and its compressed image Db data between frames from which obtains a reproduced image of high quality, in an M medium of data storage, as the compressed image data of the moving illustration. In Figure 12, the reference numbers Dl-Dk designate compressed image data corresponding to other moving illustrations which contain Hl-Hk headers respectively. The Ha header of the Da data contains a flag indicating that the Da data is suitable for use in independent playback. The header Hb of the data Db contains a flag indicating that the data Db is less suitable for use in independent reproduction. In the fast forward reproduction process, according to the respective flags contained in the corresponding He and Hb headers, the compressed image data Da within frame is read from the data storage medium M as compressed image data of a illustration in movement. On the other hand, in the normal reproduction process, the compressed image Db data between frames is read from the data storage medium M. The second solution is to insert a plurality of coded frame data pieces corresponding to the illustrations I into the compressed image data Db between frames at shorter intervals than the normal ranges. In general, the encoded frame data corresponding to the illustrations I are inserted into the compressed image data so that two of the plurality of frames reproduced for 0.5 seconds are the I images. This compressed image Db data between frames contains the flag indicating that the Db data is suitable for use in the independent reproduction. In this case, in the fast forward reproduction process, according to the image type flags (not shown) added to the frames, each indicating that the corresponding coded frame data corresponds to the illustration I, only the data encoded frames corresponding to illustrations I can be decoded. The third solution is, because the encoded frame data corresponding to any of the P illustrations are reproducible independently, add flags indicating that this or these are encoded frame data. Such encoded frame data corresponding to some of the illustrations P are obtained by coding without reference to the image data of another frame other than the encoded frame data corresponding to the illustrations I, although the corresponding type of picture flags indicate the "P illustrations". The encoded frame data corresponding to these specified P illustrations are independently reproducible. Therefore, the flags indicating that the encoded frame data corresponding to these specified P illustrations are suitable for use in independent reproduction are also added to them. Thus, in the process of rapid advancement reproduction, according to the flags of type of illustration and these flags indicating the adequacy of the independent reproduction (not shown), only encode frame data corresponding to the illustrations I and the specified P illustrations are decoded. Figure 11 (c) shows a structure of the image data compressed between frames which contains the banners of the appropriate independent previous reproduction added to the encoded frame data corresponding to the specified P illustrations. In the compressed image data Dc between frames, each of the headers Hcl, Hc2, ..., Hcf containing the flags of the appropriate are inserted just before the frame data encoded Pes (1) -Pes (f) which correspond to the specified P illustrations (expressed as P 'in the figure), respectively. In the figure, He designates a header of the compressed image data Dc between frames, and Pc (l) -Pc (n) designates frame data encoded from respective frames. The header structures of the compressed image data Da and Db will be described with reference to FIG. 13. In FIG. 13, for purposes of simplicity, the compressed image data is shown without differentiating between compressed image data Da within Frames and compressed image Db data between frames.

As previously mentioned, the compressed image data D comprises the header containing data common to the respective frames which are placed at the beginning of the data D and the following frame coding data P following. The H header is constituted by a synchronous Hsd signal, data common to the respective Hcd frames, an Hfd flag that is related to what is suitable for independent reproduction, and alignment Had data to align this data. The compressed image data corresponding to one of the moving pictures in this manner contains the information (flag) which indicates whether the encoded frame data corresponding to all the frames are independently reproducible or not. When the encoded frame data corresponding to all the frames of a moving illustration are reproducible independently, the flag has a value indicating that the corresponding compressed image data is suitable for use in independent reproduction, whereas when one of the moving illustrations contain few encoded frame data which are reproducible independently, the flag has a value indicating that the compressed image data corresponding are less suitable for use in independent reproduction. The flag that is contained in the header H includes common data (data common to respective frames) at the beginning of this compressed image data. A description of the data alignment examples in the header of the compressed image data is given below with reference to tables 1-3 shown below. The data shown in tables 1-3 are continuously aligned in the header in the order of transmission. Placed at the beginning of the header, there is a synchronous signal 902 that indicates the start of the motion illustration, which is represented as a single fixed length code (32 bits). After the synchronous signal 902, various common data types 903-913 are placed, common to the respective frames. In the common data 903-913, the data 910 is represented by a variable length code and the data 903-909 and 911-913 are each represented by a code having a plurality of fixed bit lengths. After these common data, 903-913, a flag 914 is placed which is related to what is suitable for independent reproduction and the alignment data 915. Flag 914 indicates whether the encoded frame data or frames are reproducible randomly and independently. The value of "1" of the flag indicates that all the encoded frame data of the corresponding compressed image data of the moving illustration are reproducible independently, while the value "0" indicates that most of the frame data encoded from the corresponding compressed image data are not independently reproducible. The alignment data 915 is used to align the synchronous signal 902 through the flag 914. Following the alignment data 915, data 916 and 917 are placed in relation to the coded frame data obtained by encoding image data corresponding to the respective data of the illustration in movement. Currently, these data 916 and 917 include specific data such as DCT coefficients or quantization steps according to MPEG 1, 2 and 4, although these are illustrated as a group of data in this illustrated example. It must be remembered that the header containing such common data is placed at the beginning of the compressed image data of the moving illustration. If the frame-compressed image data includes encoded frame data which is not independently reproducible and includes some independently reproducible encoded frame data (encoded frame data corresponding to illustration I) which are they distribute periodically, the effectiveness is provided by inserting common data that contain a flag in relation to the possibility of independent reproduction instead of a flag in relation to what is suitable for independent reproduction. The first flag indicates whether the corresponding coded frame data is reproducible independently or not, without reference to other frame data. In the fast forward reproduction process performed with the compressed image data between frames within which common data is periodically inserted, the frame data independently reproducible, for the illustration I are selectively decoded.

[TABLE 1] 901" 902- 903 < 904 907 [Table 2] [Table 3] 912 913 914"915-916 917 When the fast forward playback process or the fast rewind playback process is performed to the compressed image data of the illustration in movement, the encoded frame data is randomly selected from the compressed image data and then decoded, and subsequently it is necessary to quickly decide whether the compressed image data is suitable or not for use in the fast forward or the playback process. if the encoded frame data of the compressed image data is reproducible independently or not. Nevertheless, it is impossible to quickly decide this (what is suitable for independent reproduction and the possibility of independent reproduction) from the headers that are added to the conventional compressed image data and the encoded frame data. In order to decide whether the compressed image data is suitable or not for use in independent reproduction, the flag (data 914 shown in tables 1-3) in the header containing the common data is extracted and analyzed. To verify whether the value of flag 914 in the header is "1" or not, it is necessary that all common data 903-913 placed before flag 914 be extracted and then analyzed by parsing of these common data before flag 914 is analyzed. For example, until it has been verified that the value of common data 903a is "1", it is impossible to decide whether common data 903b and 903c are present.

In the header added to the conventional compressed image data, various data are placed such as the synchronous signal 902 indicating the start of the moving illustration and the common data 903-913 for the encoded frame data, before the flag which indicates that the compressed image data is suitable for use in independent reproduction. Many of these common data often serve as a switch or the like. This means that the subsequent data processing depends on the value of such common data. For the above reason, a lot of time is required when starting the data analysis of the headers until the analysis of the banner of the appropriateness of the independent reproduction is initiated. The present invention is directed to solving the problem, and an object of the present invention is to provide an image processing method comprising a coding process for producing compressed image data of a data structure which makes it possible to quickly decide whether the The compressed image data corresponds or does not correspond to a moving illustration and encodes the frame data to determine whether they are suitable for use in independent reproduction and whether they are reproducible independently or not from the aggregated headers to these data, and a decoding process in a manner adapted to the coding process. Another object of the present invention is to provide a data storage medium which contains an image processing program for performing a computer to perform the coding process and the decoding process.

DESCRIPTION OF THE INVENTION According to claim 1 of the present invention, there is provided an image transmission method for transmitting compressed image data that is obtained by compressively encoding digital image data corresponding to an image comprising a plurality of frames, and the method it comprises the steps of: transmitting a header that includes data common to the respective frames which are included in the compressed image data; and sequentially transmitting compressed image data of the respective frames which are included in the compressed image data, wherein the header includes an identification flag that indicates whether the compressed image data is suitable or not for use in a reproduction process which randomly selects compressed frame data from arbitrary frames to reproduce data from frame, and from encoded data at the beginning of the header to encode data immediately before the identification flag is set in terms of length. In the image transmission method constructed in this manner, when the added header is transmitted to the compressed image data corresponding to a moving image, the header includes the identification flag indicating whether the compressed image data is suitable or not for use in the random (independent) reproduction process for arbitrary frames, and in the header, only fixed length coding data is present before the identification flag. Therefore, when the header added to the compressed image data is analyzed, it is possible to quickly analyze the identification flag and thus decide whether the compressed image data is suitable or not for use in a random reproduction process in a short time According to claim 2 of the present invention, in the image transmission method of claim 1, the header comprises a synchronous signal indicating a header of the compressed image data, fixed length coding data and coding data. of variable length, such as the data common to the respective frames, and the flag of identification, and in the stage of transmission of the header, the flag of identification after the synchronous signal and before the variable length coding data. In the image transmission method constructed in this way, the identification flag is transmitted after the synchronous signal and before the variable length coding data. Therefore, at the end of the decoding, the identification flag is analyzed immediately after the synchronous signal is analyzed. According to claim 3 of the present invention, there is provided an image processing method for compressively encoding digital image data corresponding to an image comprising a plurality of frames, and the method comprises the steps of: generating a header which includes data common to the respective frames; and compressively coding frame data corresponding to the respective frames to produce compressed frame data, wherein the header is generated in such a way that it includes an identification flag indicating whether the compressed image data is suitable or not for use in a random reproduction process which randomly selects frame data compressed from arbitrary frames to reproduce frame data, and from encoding data at the beginning of the header to encode data immediately before the identification flag is set to length.

In the image processing method constructed in this manner, when digital image data corresponding to a moving illustration is compressively encoded to produce compressed image data, the header including the identification flag indicates whether the compressed image data they are suitable or not for use in the random (independent) reproduction process, and are added to the compressed data, and in the header, from the coding data at the beginning of the data to the encoded data just before it is set the identification flag in length. Therefore, when the header is analyzed, it is possible to quickly analyze the identification flag and thus decide whether the compressed image data is suitable or not for use in the random reproduction process in a short time. According to claim 4 of the present invention, in the image processing method of claim 3, the step of compressively coding frame data is performed after the step of generating the header comprises generating a synchronous signal indicating a header of the compressed image data, the identification flag and the data common to the respective frames, in this order. In the image processing method built in this way, in the header, the synchronous signal, the Identification flag and common data are aligned in this order. Therefore, at the decoding end, the identification flag is analyzed immediately after the synchronous signal is analyzed. According to claim 5 of the present invention, in the image processing method of claim 3, the step of compressively coding frame data is performed after the step of generating the header, and the step of generating the header comprises generating a synchronous signal indicating a header of the compressed image data, fixed-length code data such as the data common to the respective frames, the identification flag and the variable-length code data, such as the data common to the frames respective, in this order. In the image processing method constructed in this manner, in the header, the synchronous signal, the common data of the fixed-length code, the identification flag and the variable-length code data are aligned in this order. Therefore, the identification flag is analyzed immediately after the synchronous signal is analyzed without analyzing the common data as required. According to claim 6 of the present invention, in the image processing method of claim 3, the step of compressively encoding data frame comprises: a first step of compressive coding to compressively encode frame data corresponding to a frame to be processed without reference to the frame data corresponding to another frame, to produce first compressed frame data; and a second compressive coding step for compressively encoding frame data corresponding to a frame to be processed with reference to frame data corresponding to another frame, to produce second compressed frame data, wherein the identification flag included in the compressed image data constituted from the first compressed frame data indicates that the compressed image data is suitable for use in the random production process, and the identification flag included in the compressed image data composed of the first data of frame compressed and the second compressed frame data indicates that the compressed image data is less suitable for use in the random reproduction process. In the image processing method constructed in this way, the identification flag of the compressed image data obtained by the first compressive coding process without reference to another frame indicates that the compressed image data is suitable for use in the process of random reproduction, while the flag identifying the compressed image data obtained by the second compressive coding process with reference to another frame and to the first compressive coding process indicate that the compressed image data is less suitable for use in the random reproduction process. Therefore, at the decoding end, it is possible to quickly distinguish between the compressed image data suitable for use in the random reproduction process and the compressed image data with higher coding efficiency which are less suitable for use in the process of random reproduction. According to claim 7 of the present invention, the method of image processing of claim 3 further comprises the step of: generating an auxiliary header including data common to the respective frames and individual data for a specified frame, wherein generates an auxiliary header such that the auxiliary header is added before the compressed frame data of the specified frame when performing the compressively coding stage of frame data after the header generation step, and in the generating step the auxiliary header, the auxiliary header is generated such that it includes a flag indicating whether the compressed frame data of the specified frame is independently reproducible or not, without reference to frame data of another frame, and from code data to the principle of the auxiliary header to encode data immediately before the flag is set in terms of length. In the image processing method constructed in this manner, the auxiliary header is added to compressed frame data of the specified frame of the compressed image data, and the auxiliary header includes the flag indicating whether the frame data compressed from the specified frame they are reproducible independently or not. In addition, in the auxiliary header from the code data at the beginning of the same to encode data just before it is fixed in terms of length the identification flag. Therefore, it is possible to decide whether the compressed frame data is reproducible independently or not for each frame. According to claim 8 of the present invention, there is provided an image processing method for decoding compressed image data contained by compressively encoding digital image data corresponding to an image comprising a plurality of frames to provide reproduced image data. corresponding to the image, and the method comprises the steps of: analyzing a header including data common to respective frames which are included in the compressed image data; and decoding compressed frame data obtained by compressively encoding frames frame data respective and included in the compressed image data to provide reproduced frame data, wherein the step of analyzing the header comprises analyzing t fixed-length code data from encoding data at the beginning of the header to encode data immediately before that an identification flag is included in the header and that it indicates whether the compressed image data is suitable or not for use in a random production process which randomly selects data from arbitrary frame compressed frames to provide reproduced frame data, and then analyze the identification flag. In the image processing method constructed in this manner, when the compressed image data corresponding to a moving illustration is decoded, the header is analyzed so that the analysis of the fixed-length code data from the beginning of the they are followed by analysis of the identification flag which indicates whether the compressed image data is suitable or not for use in independent reproduction. Therefore, when the header is analyzed, it is possible to quickly analyze the identification flag and thus decide whether the compressed image data is suitable or not for use in the process of random reproduction in a short time.

According to claim 9 of the present invention, in the image processing method of claim 8, the step of decoding compressed frame data is performed after the step of analyzing the header, and the step of analyzing the header comprises analyzing a synchronous signal indicating a header of the compressed image data, the identification flag and the data common to the respective frames, in this order. In the image processing method constructed in this way, when the header is analyzed, the synchronous signal, the identification flag and the common data are analyzed in this order. Therefore, in the decoding process, the identification flag is analyzed immediately after the synchronous signal is analyzed. According to claim 10 of the present invention, in the image processing method of claim 8, the decoding step of compressed frame data is performed after the step of analyzing the header, and the step of analyzing the header comprises analyzing a synchronous signal indicating a header of the compressed image data, fixed-length code data as data common to the respective frames, the identification flag and the variable length code data as the data common to the respective frames , in this order.

In the image processing method constructed in this way, when the header is analyzed, the synchronous signal, the fixed-length code data as the common data, the identification flag and the variable-length code data as the common data they are analyzed in this order. Therefore, the identification flag is analyzed immediately after the synchronous signal is analyzed, without analyzing the common data, as required. According to claim 11 of the present invention, in the image processing method of claim 8, the step of analyzing the header and the step of decoding compressed frame data is performed for first compressed image data consisting of first data. of compressed frame obtained by compressively encoding frame data of a frame to be processed without reference to first frame data of another frame, and is performed for second compressed image data consisting of the first compressed frame data and second frame data. frame compresses obtained by compressively encoding frame data of a frame to be processed with reference to frame data of another frame, and the random reproduction process is performed only for the first compressed image data according to the identification flag .

In the image processing method constructed in this way, reproduction is performed for each of the compressed image data that is obtained by the first compressive coding process without reference to another frame and compressed image data that is obtained by the second compressive coding process with reference to another framework and to the first compressive coding process. In addition, only the first compressed image data are reproduced randomly. Therefore, at the decoding end, the random reproduction process is preferably performed. According to claim 12 of the present invention, the image processing method of claim 8 further comprises the step of: analyzing an auxiliary header added to the compressed frame data of a specified frame and including data common to the respective frames and individual data for the specified frame, in which the auxiliary header is analyzed for the specified frame when performing the stage of decoding compressed frame data after the stage of analyzing the header, and the stage of analyzing the auxiliary header it comprises analyzing fixed-length code data from code data at the beginning thereof to encode data immediately before a flag is included in it indicating whether the frame data Tablets of the specified frame are reproducible independently or not, without reference to the frame data of another frame, and then analyze the flag. In the image processing method constructed in this way, the auxiliary header aggregated to the compressed frame data of the specified frame is analyzed, so that the fixed length code data of the same frame is analyzed and then the flag in the auxiliary header. Therefore, it is possible to decide whether the compressed frame data is reproducible independently or not for each frame. According to claim 13, of the present invention, an image processing apparatus for compressively encoding digital image data corresponding to an image comprising a plurality of frames for producing compressed image data is provided, and the apparatus comprises: a prediction data generator for generating prediction framework data for objective framework data corresponding to a framework to be processed based on the objective framework data; a calculation means for transmitting either the difference frame data as a difference value between the target frame data and the prediction frame data, or the target frame data according to a control signal; a data compressor to compress the data output from the calculation means to produce compressed data; a variable length encoder for performing variable length coding of the compressed data transmitted from the data compressor and transmitting the compressed frame data of each frame; and a control means for generating a header that includes data common to the respective frames based on the digital image data, and controlling the computing means according to the identification flag indicating whether the compressed image data is suitable for use or not in a random reproduction process, which randomly selects compressed frame data from arbitrary frames to reproduce frame data, wherein the variable length encoder transmits the header including the identification flag, in which the code data at the beginning of the header to encode data immediately before the identification flag is set to length. In the image processing apparatus constructed in this way, when digital image data corresponding to a moving illustration is compressively encoded to provide compressed image data, the header includes the identification flag indicating whether the compressed image data is suitable or not for use in random (independent) reproduction for arbitrary frames and that is added to the compressed image data and, in the header, from code data at the beginning of the same to encode data just before the identification flag is fixed in length. Therefore, when the header added to the compressed image data is analyzed, it is possible to quickly analyze the identification flag and thus decide whether the compressed image data is suitable or not for use in the random reproduction process, in a short time According to claim 14 of the present invention, in the image processing apparatus of claim 13, the variable length encoder transmits the header before transmitting the compressed frame data of the respective frames in such a way as to transmit a synchronous signal indicating a header of the compressed image data, the identification flag and the data common to the respective frames, in this order. In the image processing apparatus constructed in this way, in the header, the synchronous signal, the identification flag and the common data are aligned, in this order. Therefore, at the decoding end, the identification flag is analyzed immediately after the synchronous signal is analyzed. According to claim 15 of the present invention, in the image processing apparatus of claim 13, the variable length encoder transmits the header before the data is transmitted. of compressed frames of the respective frames such that a synchronous signal is transmitted indicating a header of the compressed image data, fixed length coding data such as data common to the respective frames, identification flag and code data of variable length as the data common to the respective frames, in this order. In the image processing apparatus constructed in this way, in the header, the synchronous signal, the fixed-length code data as the common data, the identification flag and the variable length code data are aligned, in this order . Therefore, the identification flag is analyzed immediately after the synchronous signal is analyzed, without analyzing the common data, as necessary. According to claim 16 of the present invention, in the image processing apparatus of claim 13, the control means controls the computing means so that all frames of the image are subjected to a first compressive coding process. in which the means of calculation transmits the objective frame data, and the data compressor compresses the frame data of a frame to be processed without reference to the frame data of another frame, and the variable length encoder transmits the first compressed frame data, when the flag of identification indicates that the compressed image data is suitable for use in the random reproduction process, and the control means controls the calculation means so that the specified data of the image is subjected to a second compressive coding process in which the calculation means transmits the difference frame data, the data compressor compresses frame data of a frame to be processed with reference to the frame data of another frame, and the variable length encoder transmits the second frame data. frame compresses, and frames other than the specified frames are subjected to a first compressive coding process, when the identification flag indicates that the compressed image data is less suitable for use in a random reproduction process. In the image processing apparatus constructed in this way, the identification flag of the compressed image data which is obtained by the first compressive coding process with reference to another frame, indicates that the compressed image data is suitable for use in the process of random reproduction, while the identification flag of the compressed image data that is obtained by the second compressive coding process with reference to another frame and the first coding process, indicates that the compressed image data is less suitable for use in the reproduction process random Therefore, at the deepening end, it is possible to quickly differentiate between the compressed image data suitable for use in the random reproduction process and the compressed image data with higher coding efficiency which are less suitable for use in the process of random reproduction. According to claim 17 of the present invention, an image processing apparatus for decoding compressed image data that is obtained by compressively encoding digital image data corresponding to a frame comprising a plurality of frames to provide an image is provided. reproduced that corresponds to the image, and the apparatus comprises: an analyzer for analyzing a header included in the compressed image data to generate header information, and analyzing the data corresponding to each frame included in the compressed image data and transmitting the data of frame tablets; a data decompressor for decompressing the compressed frame data to produce uncompressed frame data; a calculating means for transmitting either frame obtained by adding the decompressed frame data and the prediction frame data thereof, or the frame data decompressed as frame data reproduced, according to a control signal; a prediction data generator to generate prediction framework data for a framework that is going to be processed from objective framework data to be decompressed that correspond to the framework to be processed; and a control means for controlling the calculation means according to an identification flag included in the header information and indicating whether the compressed image data is suitable or not for use in a random reproduction process which randomly selects data of frame frames of arbitrary frames to reproduce frame data, where the analyzer analyzes the header in such a way that it analyzes the identification flag without analyzing the common data composed of fixed-length code data from code data in the beginning of the header to encode data immediately before the identification flag, as required. In the image processing apparatus constructed in this way, when compressed image data corresponding to a moving illustration is decoded, the header is analyzed so that the identification flag indicates whether the compressed image data is suitable or not. for use in the random reproduction process, and analyzed without analyzing the fixed length code data of the start thereof. Therefore, when the header added to the compressed image data is analyzed, it is possible to quickly analyze the identification flag and thus decide whether the image data compri idoe eon suitable or not for the process of random reproduction in a short time. According to claim 18 of the present invention, in the image processing apparatus of claim 17, the analyzer analyzes the header in such a way that it analyzes a synchronous signal indicating a header of the compressed image data, the flag of identification and data common to the respective data, in this order, according to the order in which these data are entered into the analyzer. In the image processing apparatus constructed in this way, the synchronous signal, the identification flag and the common data are analyzed, in this order. Therefore, at the decoding end, the identification flag is analyzed immediately after the synchronous signal is analyzed. According to claim 19 of the present invention, in the image processing apparatus of claim 17, the analyzer analyzes the header in such a way that it analyzes a synchronous signal indicating a header of the data of the compressed image, the data of fixed-length code such as the data common to the respective frames, the identification flag and the variable-length code data, such as the data common to frames respective, in this order, according to the order in which these data are entered into the analyzer. In the image processing apparatus constructed in this manner, the synchronous signal, the fixed-length code data such as the common data, the identification flag and the variable length code data are analyzed as the common data, in this order. . Therefore, the identification flag is analyzed immediately after the synchronous signal is analyzed, without analyzing the fixed-length code data as common data, as required. According to claim 20 of the present invention, in the image processing apparatus of claim 17, the control means controls the computing means so that all frames of the image are subjected to a first decompressive decoding process. without reference to another frame in which the decompressed frame data corresponding to the frame to be processed is transmitted from the calculation means as reproduced frame data, when the identification flag indicates that the compressed image data is suitable for use in the random reproduction process, and the control means controls the calculation means so that the specified frames of the image are subjected to a second decompressive decoding process with reference to another frame in which is a value of addition of the decompressed frame data of the frame to be processed and frame data reproduced from another frame, from the calculation means as frame data reproduced from the same frame that is going to be processed and frames other than the specified frames that undergo the first decompressive decoding process, when the identification flag indicates that the compressed image data is less suitable for use in the random production process. In the image processing apparatus constructed in this way, the reproduction is performed for each of the compressed image data that is obtained by the first compressive coding process without reference to another frame and the compressed image data obtained by the second one. compressive coding process with reference to another framework and the first compressive coding process. In addition, only the first compressed image data are reproduced randomly. Therefore, at the decoding end, the random reproduction process is preferably performed. According to claim 21 of the present invention, a data storage means for storing an image processing program for compressively encoding digital image data corresponding to an image comprising a plurality of images is provided. framework, and the image processing program allows a computer to perform a compressive coding process to the digital image data according to an image processing method of claim 3. The use of the data storage medium constructed from this Thus, when loading the coding program into the computer, when the digital image data corresponds to a moving illustration is compressively encoded to provide compressed image data, the header includes an identification flag that indicates whether the compressed image data is suitable or not for use in random (independent) reproduction, which is added to the compressed image data, and in the header, from the data encoded at the beginning of the code data before it is fixed in length. identification flag. Such a coding process is executed by the computer. Therefore, it is possible to implement an image coding process by programming elements (software) in which the identification flag is analyzed quickly and in this way it is determined whether the compressed image data is suitable or not for use in a Random playback process in a short time when the header added to the compressed image data is analyzed.

According to claim 22 of the present invention, a data storage means is provided for storing an image processing program for decoding uncompressed compressed image data that is obtained by compressively encoding digital image data corresponding to an image. comprising a plurality of frames, and the image processing program allows a computer to perform a decoding process for the compressed image data according to an image processing method of claim 8. The use of the constructed storage medium in this way, when loading the decoding program on the computer, when compressed image data corresponding to a moving illustration is encoded, the header is analyzed so that the analysis of the fixed-length code data from the beginning of the same is followed by analysis of the flag of ident ification that indicates whether the compressed image data is suitable or not for use in independent reproduction. Such a decoding process is executed by the computer. Therefore, it is possible to implement the decoding process by programming element (software) in which the identification flag is analyzed quickly and in this way it is decided if the compressed image data is suitable or not for use in the process Random playback in a short time when analyzing the header added to the compressed image data.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a diagram showing the data structures of the image signals according to a first embodiment, in which Fig. 1 (a) shows a sequence header, Fig. 1 (b) shows compressed image data. of a data structure which is suitable for use in independent reproduction, and figure 1 (c) compressed image data of a data structure with high compression efficiency. Fig. 2 is a flow diagram for explaining an image coding process performed by an image processing method of the first embodiment. Figure 3 is a block diagram showing a structure of an image processing apparatus which performs the image coding process of the first embodiment. Figure 4 is a diagram showing a data structure of the compressed image data, which is different from the compressed image data of the first embodiment.

Fig. 5 is a diagram showing data structures of the compressed image data according to a modification of the first embodiment, wherein Fig. 5 (a) shows a data structure of the compressed image data which is suitable for use in independent reproduction, and Figure 5 (b) shows a data structure of compressed image data with a high compression efficiency which is suitable for use in independent reproduction. Fig. 6 is a flow chart explaining an image coding process performed by the image processing method according to the modification of the first mode. Fig. 7 is a flow diagram for explaining an image decoding process according to an image processing method according to a second embodiment of the present invention. Figure 8 is a block diagram showing an image processing apparatus which performs a decoding process of the second embodiment. Figure 9 is a diagram showing a data storage medium (Figure 9 (a) and Figure 9 (b)), which contains programs that allow a computer system to perform the coding process and the process of decoding of each mode, and Figure 9 (c) is a diagram showing the computer system. Fig. 10 is a diagram showing the data structures of conventional coded image signals (compressed data), wherein Figure 10 (a) shows compressed image data corresponding to a moving illustration, Figure 10 (b) shows compressed image data which is suitable for use in independent reproduction, and Figure 10 (c) shows compressed image data with high compression efficiency. Figure 11 is a diagram for explaining the problems associated with conventional coded image signal data structures, wherein Figure 11 (a) shows compressed image data which are suitable for use in stand-alone reproduction, Figure 11 ( b) shows compressed image data with high compression efficiency, and FIG. 11 (c) shows compressed image data with high compression efficiency which are suitable for independent reproduction. Figure 12 is a conceptual diagram for explaining the data storage medium which contains the compressed image data corresponding to the various types of moving illustrations.

Figure 13 is a diagram showing a sequence header that is included in the conventional compressed image data.

BEST MODE FOR CARRYING OUT THE INVENTION The preferred embodiments of the present invention will now be described.

[Mode 1] An image transmission method of the first embodiment is capable of transmitting compressed image data that is obtained by compressively encoding digital image data (moving image data) that corresponds to a moving illustration comprising a plurality of frames. Assuming that the digital image data used in this mode is pre-recorded data as a digital image signal comprising a luminance signal and a chrominance signal. In addition, the digital image signal may be an image signal corresponding to an image in a conventional image space (display image) of a rectangular shape, or it may be an image signal corresponding to an object region (VOP). : video object plane) that includes an object (an image of an arbitrary shape) in the exhibition image. Figure 1 is a diagram showing the data structures of the image signals according to the first embodiment, wherein Figure 1 (a) shows a data structure of a header of the compressed image data. The compressed image data Dv comprises encoded frame Dp data of the respective frames and an Hv header added before the frame data Dp encoded, and which indicates the attribute of the data Dp. The encoded frame data Dp is obtained by compressively encoding image data corresponding to the respective frames of a moving picture. The header Hv comprises a synchronous Hsd signal indicating the start of the compressed image data Dv, an identification flag Hfd (flag RA) indicating whether the compressed image data Dv is suitable for use in random reproduction, Hcd data (data common) which are common to the respective frames, and alignment data Had to align the synchronous Hsd signal, the identification flag Hfd and the common Hcd data. In random reproduction, the encoded frame data corresponding to arbitrary frames are selected and reproduced arbitrarily.

As already described in the prior art, the compressive coding process includes two types of processing, specifically the intra-frame coding process and the frame-to-frame coding process. Returning to Figure 1 (b), the first compressed image Dva data is obtained by performing the coding process in frame for the frame data P (l) -P (n) (see Figure 10 (a)) that correeponde a todoe frames the D data of moving image and is suitable for use in random reproduction. Therefore, the value of the identification flag Hfd, in a header Hva of the first compressed image data Dva, is "1", which indicates that the compressed image data Dva is suitable for use in random reproduction. Returning to FIG. 1 (c), the second compressed image data Dvb is obtained by performing the frame coding process to the frame data P (l) corresponding to a frame at the start of the illustration data D in FIG. movement and when performing the coding process between frames to the following frame data P (2) -P (n) and is less suitable for use in random reproduction. Therefore, the value of the identification flag Hfd 0 in the header Hvb of the second compressed image data Dvb is "0", which indicates that the compressed image data Dvb is less suitable for use in random reproduction.

In the transmission method of the first mode, the header including the common data is transmitted, and then the compressed frame data of the respective frames are transmitted sequentially. When the Hv header is transmitted, the synchronous Hsd signal, the identification Hfd flag, the common Hcd data and the alignment Al data are transmitted in this order. Subsequently, an image processing method of the first embodiment will be described. In this method of image processing, the data Digital image D (see Figure 10 (a)) is compressively encoded to produce one of the first compressed image data Dva and the second compressed image Dvb data as the compressed image Dv data. Figure 2 is a diagram for explaining the image processing method of the first embodiment showing a flow of the coding process according to the processing method. The coding process is started (step 101). The synchronous sequence signal Hsd, which indicates the start of the compressed image data Dv, is produced (step 102). The synchronous Hsd signal is represented by a unique code (32 bits). An identification flag Hfd code (step 103) is produced. The Hfd flag indicates whether all the frames of the Digital motion detection is encoded or not compressively without reference to a different frame to the target frame that is to be processed. The value of the Hfd flag is "1" when all the frames are to be compressively encoded without reference to another frame, and the value is "0" when it will not happen in this way. A code of the common data necessary to reproduce the compressed image Dv data at one reproduction end and a code of the alignment data and the like is produced (step 104). The data (frame data) corresponding to the respective frames are compressively encoded sequentially. To be specific, when entering the frame data P (l), which corresponds to the first frame (step 105), the frame data is compressively encoded to produce frame Dp data encoded according to the value of the Hfd flag of identification (step 106). Step 106 will be briefly described below. When the value of the identification flag Hfd is "1", all data P (l) -P (n) of marcoe eon encoded within the frame. To be specific, the image data of each frame is divided into a plurality of macroblocks (image areas) each consisting of (16 x 16) pixels in a framework. In addition, the image data of each macroblock are divided into sub-blocks corresponding respectively to image spaces each consisting of (8 x 8) pixels. The image data of each sub-block is subjected to the DCT process to be transformed into DCT coefficients for each sub-block, which are quantized to generate quantization coefficients which are converted into a variable length code. The macroblocks that constitute a frame in this way are processed and the resulting encoded frame data is transmitted. On the other hand, when the value of the identification flag Hsd is "0", the data P (l) is encoded within the frame, and the other frame data P (2) -P (n) are encoded between the frames . The coding process between frames will be briefly described below. Initially, by means of the method for detecting movement of an image in a frame such as "block matching", a prediction macroblock is detected. It is detected as the prediction macroblock a region consisting of (16 x 16) pixels with the smallest errors between pixel values thereof and the pixel values of a target macroblock to be encoded, from image data what they correspond to an encoded frame which is temporarily adjacent to a frame that is to be encoded. Subsequently, the image data of the prediction macroblock is subtracted from the image data of the target macroblock to produce difference data of the target macroblock, which is divided into difference signals corresponding respectively to subblock with each of (8 8) pixels. Then, the difference signals are submitted to the DCT process for each sub-block, in order to generate the DCT coefficients, which are quantized for each sub-block to generate quantization coefficients, which are converted into variable length code. The macroblocks that constitute a frame in this way are processed and the resulting encoded frame data is processed. Subsequently, in step 106 it is decided whether the input frame data is the last frame or not (step 107).

When it is decided that the input frame data is not the last frame, step 106 is performed again, or otherwise, the coding process is completed (step 108). In this way the coding process is performed to produce compressed image data Dva shown in Fig. 1 (b) which are suitable in random reproduction, or compressed image data which are less suitable for the random reproduction shown in Figure 1 (c). The compressed image data is transmitted to a decoding end by means of a communication line or stored in a storage medium and then supplied to it.

The following tables 4, 5 and 6 show an example of a structure of the compressed image data produced in this way, in particular, header data alignment. These tables show the structure of the compressed image Dv data shown in Figure 11 (a). That is, these tables do not differentiate between the compressed image Dva data in Fig. 1 (b) and the compressed image Dvd data in Fig. 1 (c).

[Table 4] [Table 5] [Table 6] The data included in Tables 4-6 are continuously aligned in the order of transmission.

At the beginning of the header, a synchronous 802 signal is placed indicating the start of the moving illustration. Synchronous signal 802 is represented by a fixed length code (32 bits). Subsequent to the synchronous signal 802, the datoe 814 corresponding to a 1-bit identification flag Hfd are placed. After the data 814, data (common data) 803-813 common to the respective frames is placed. In the common data 803-813, the data 803-809 and 811-813 are each represented by a code having a plurality of fixed bit lengths and the data 810 is represented by a variable length code. Subsequent to the common data 803-813, alignment data 815 is placed, which aligns the synchronous signal 802, the identification flag code 814, and the common 803-813 data. After the alignment data 815, the data 816 and 817 are placed in relation to the coding frame obtained by carrying out the coding process within frames or the coding process between frames for frames of the moving illustration. At present, these data 816 and 817 include DCT coefficients or quantization steps according to MPEG 1, 2 and 4, although these are shown as a group of data in this illustrated example. Although the data 814 corresponding to the identification flag Hfd is placed just after the signal synchronous sequence (data 802), can be placed after data 803 (N separate bits of data 802). Preferably, the data 814 is placed before the fixed length code data with a variable length code data decision condition. In any case, it is desirable to place the data 814 near the head of the common data. Next, a description will be provided of a processing apparatus (image coding apparatus) which performs the compressive coding process according to the image processing method of the first embodiment. Figure 3 is a block diagram showing the image coding apparatus of the first embodiment. An image coding apparatus 100a is adapted to perform compressive coding of digital image data (moving illustration data) which corresponds to a moving image comprising a plurality of frames for producing compressed image data. The image coding apparatus 100a includes a prediction data generator 406 for generating frame data 420 of predicting the objective frame data corresponding to a frame to be processed, based on the objective frame data, and an adder 402 for transmitting difference frame data as a difference value between the objective frame data 416 and the prediction frame data 420. The image coding apparatus 100a further includes a data compressor 403 for compressing output data 421 of the adder to produce data 423 compressed, and a variable length encoder 414 for performing variable length coding of the compressed data 423 that is transmitted from the compressor 403. The data compressor 403 comprises a 404 DCT unit for performing the DCT process to the output data 421 of the adder 402 and a quantizer 405 for quantifying the output data 422 of the 404 DCT unit and transmitting the data 423 tablets In the image coding apparatus 100a, the moving picture data 416 is input to a first input terminal 401 and supplied to the prediction data generator 406 via a first switch 434a, while the frame data 420. of prediction are supplied to the adder 402 through a second switch 434b. The compressed data 423 is transmitted from the data compressor 403 and supplied to the prediction data generator 406 via a third switch 434c and motion information (motion vector) 418 generated by the prediction data generator 406 which it is transmitted to the variable length encoder 414 through a fourth switch 434d.

The image coding apparatus 100a further includes a controller 433 for generating header information 436 which includes data common to respective frames based on the digital image data and transmitting the header information 436 to the variable length encoder 414, and a control of On-off of the switches 434a-434d through the use of control signals 437a-437d, respectively. An externally inputted control signal 435 contains the identification flag Hfd which indicates whether the compressed image data is suitable or not for use in random reproduction in which the compressed frame data corresponding to an arbitrary frame is selected and reproduced. The variable length encoder 414 is used to perform varying length encoding of header information 436, motion information 418, and compressed data 423, and transmits a stream 431 of bits such as compressed image data Dv to a terminal 415 output. The variable length encoder 414 transmits the header Hv according to the header information 413. In the Hv header, from the start code through the code just before the code length of the identification flag is set. The construction of the prediction data generator 406 will be described.

The prediction data generator 406 includes a data decompressor 407 which receives the compressed data 423 from the data compressor 403 through the third switch 434c, decompresses the compressed data 423 and transmit uncompressed data 426, and a second adder 409 which adds the data 426 decompressed to the data 420 of the prediction frame and transmits 427 data reproduced. Data decompiler 407 comprises an inverse quantizer 407a which inversely quantizes the compressed data 423 and a 407b IDCT unit (reverse discrete cosine transform) which submits the output data 425 to the inverse quantizer 407a to an IDCT process when transforming data into a frequency domain in data in a spatial domain and transmits the decompressed data 426. The prediction data generator 406 further includes a frame memory 410 which contains an output (reproduced data) 427 of the second adder 409 as reference image data for a frame to be processed subsequently. The frame memory 410 is used to transmit data in accordance with the externally input read signal 428. The prediction data generator 406 further includes a motion detector 411 which finds a vector MV of movement of the target block of a frame to be processed, based on the digital image data 416 of input, and transmits the motion vector MV and an address generator 412 which generates the read address signal 428 of the frame memory 410, according to the motion vector MV 418 from the motion detector 411 and a prediction signal reading unit 413 which reads data in an area in the frame memory 410 which is specified by the read address signal 428 and transmits the prediction frame data 420. Subsequently, the operation will be described. The digital image data is input to the first input terminal 401, while information of the identification flag Hfd (flag information) 435 is input to the second input terminal 432. According to the flag information 435, the controller 433 generates control signals 437a-37d, according to which the respective on-off switches 434a-434d are controlled. The control signals 437a-437d are the same signals. When the identification flag Hfd indicates that the compressed image data Dv is suitable for use in the random reproduction, that is, its value is "1", the switches 434a-434d are deactivated, according to the control signals 437a- 437d of the controller 433, respectively. In this way, the frame data that correspond to all frames of the digital input image are encoded within the frame. The digital image data passes to the adder 402 and is sent to the data compressor 403, which performs the data compilation process (DCT process and quantization process) to the digital image data, in accordance with the MPEG standard. The compressed data (quantization coefficients) 423 transmitted from the data compressor 403 is transformed into a variable length code by the variable length encoder 414. The coded frame data Pa (l) -Pa (n) corresponding to the respective frames are produced in this way. At this time, the variable length encoder 414 transforms the synchronous sequence signal (synchronous signal) Hsd, the identification flag (the value = 1), the common Hsd data, and the alignment data Had in the corresponding codes, for form the Hva header. The header Hva is formed in such a way that the synchronous Hsd signal, the identification flag Hfdl, the common Hcd data and the alignment data Had are transmitted in this order. The header Hva is added to the coded frame data Pa (l) -Pa (n) and the resulting compressed image data Dva is transmitted from the variable length encoder 414.

On the other hand when the identification flag Hfd indicates that the compressed image Dv data is less suitable for use in random reproduction, that is, its value is "0", the controls 434a-434d are controlled by on-off according to with control signals 437a-437d from controller 433, respectively. In this way, for example, the encoded frame data P (l) corresponding to the first frame are encoded within the frame, and the other frame data P (2) -P (n) corresponding to the following frames are encoded between the frames. The coding process within the framework is as described above, and therefore, the coding process between frames will be described below. When the switches 434a-434d are inactivated according to the control signals 437a-434d of the controller 433, the digital image data inputted is encoded between the frames. The prediction data generator 406 generates prediction data 420 for the frame to be processed based on the encoded frame data. The first adder 402 subtracts the prediction frame data 420 from the frame data 416 corresponding to the frame to be processed to generate difference frame data 421. In the data compressor 403, the 404 DCT unit transforms the difference frame data 421 into data 422 in the data domain. frequency and in addition, quantifier 405 quantifies the data 422 in the quantification coefficients. The resulting compressed data 423 is transmitted to the variable length encoder 414. The variable length encoder 414 transforms the compressed data (quantization coefficients) 423 into variable length codes to produce the encoded frame data Pb (2) -Pb (n). The compressed data 423 is also input to the prediction data generator 406 via the third switch 434c. In the data decompiler 407, the inverse quantizer 407a inversely quantizes the data 423 tablets in the data 425 in the frequency domain, and the reverse IDCT unit 407b transforms the data 425 into the data 426 in the spatial domain and transmits the restored data. The second add-on 409 adds the restored data 426 to the prediction data 420 and transmits the reproduced data 427, which is stored in the frame memory 410 as reference data used to encode a sub-frame. The motion detector 411 detects the movement information of the image between the frames by matching the blog or the like based on the input digital image data 416, and transmits the motion vector 418 to the 412 steering generator. The address generator 412 generates the address signal 428 to specify a memory area in the frame memory 410, according to the motion vector 418. The prediction signal reading unit 413 reads the data in the memory area of the frame memory 410 specified by the direction signal 428 as prediction data 420 and transmits the prediction data 420 to the addendor 402. The vector 418 of movement is sent to the variable length encoder 414 through the fourth switch 434d, which is to be transformed into variable length code by it. In this way, the coded frame data Pb (l) -Pb (n) is produced. At this time, the variable length encoder 414 transforms the synchronous sequence signal (synchronous signal Hsd), the identification flag Hfd2 (value = 0), the common Hcd data, the alignment data and the like into corresponding codes to form the Hvb header The Hvb header is formed so that the synchronous Hsd signal, the identification flag Hfd2, the common Hcd data and the alignment data Had are transmitted in this order. The Hvb header is added to the encoded frame data Pb (l) -Pb (n) and the compressed image Dvb data resulting are transmitted from the variable length encoder 414. Thus, according to the first embodiment, when the digital image data corresponding to one of the moving pictures to compressively produce compressed image data is compressively encoded, the identification flag indicates whether the compressed image data is suitable. or not for use in random (independent) playback are added just after the synchronous Hsd signal is placed at the beginning of the header. Therefore, it is possible to quickly analyze the identification flag when analyzing the header added to the compressed image data corresponding to an image, and in this way to decide whether the compressed image data is suitable or not for use in reproduction randomly in a short time. Although in the first embodiment the identification flag is placed just after the synchronous Hsd signal in the header Hv of the compressed image data Dv, the structure of the header is not limited to this. For example, just like the Hvm header shown in Figure 4, the synchronous Hsd signal, the first common Hcdl data represented as a fixed length code, an identification Hfd flag, second common Hcd2 data represented as a variable length code , and the alignment data Had can be aligned in this order.

[MODIFICATION OF THE FIRST MODALITY] Figures 5 and 6 are diagrams for explaining an image transmission method and an image processing method according to a modification of the first embodiment, respectively. In Fig. 5 (a), first compressed image Dvc data values suitable for use in random reproduction are shown, and in Fig. 5 (b) second compressed image Dvd data are shown, less suitable for use in random reproduction. . The first compressed image Dvc data is obtained by performing the coding process within the frame for the frame data P (l) -P (n) (see FIG. 10 (a)) of the respective frame of the illustration data D moving. In the compressed image Dvc data, the frame headers ((VLO) Hvc (1) -Hvc (n) are added before the coded frame Pa (l) -Pa (n) data, respectively. Hvc (1) headers -Hvc (n) contains an Hfdl flag that indicates whether each of the corresponding Pa encoded frame data Pa (l) -Pa (n) are reproducible or not independently (randomly). Pa (l) -Pa (n) of encoded frames are obtained by coding process within the frame, the value of the Hfdl flag in each of the headers Hvc (1) -Hvc (n) is §1", which indicates that the corresponding encoded frame data is randomly reproducible.In each of the headers Hvc (1) -Hvc (n), like the Hv header included in the first compressed image Dv data shown in Figure 1, the synchronous Hsd signal, the Hfdl flag, the common Hdc data and the alignment Al data are aligned, in this order. to the compressed image Dvc data the header including the identification flag Hfd indicating that the compressed image Dvc data is suitable for use in random reproduction, although this is not shown.The second compressed image Dvd data is obtained at the perform the coding process within the framework for frame data that correspond to frames specified for moving illustration data D and when performing the coding process between frames for other frames. Compressed Dvd data, such as compressed Dvc image data, is added to the headers (VOL) Hvd (l), ..., Hvd (r), ..., Hvd (n) before encoded frame data Pd (l), ... Pd (r), ... Pd (n), respectively, each of the frame headers (VOL) Hvd (l), ..., Hvd (r), ..., Hvd (n) contains an Hfdl flag or an HfdO flag that indicates if each of the corresponding encoded frame data Pd (l), ... Pd (r), ... Pd (n) is randomly reproducible or not. Since the encoded frame data Pd (l) and Pd (r) is obtained by the coding process within the frame, the value of the Hfdl flag in each of the Hvd (l) and Hvd (r) headers is 3 -jl ", which indicates that each of the encoded frame data Pd (l) and Pd (r) is randomly reproducible.On the other hand, the encoded frame data Pd (2), Pd (3), Pd ( 4), ..., Pd (nl), Pd (n) are obtained by the process of coding between frames, the value of the HfdO flag in each of the headers Hvd (2), Hvd (3), Hvd ( 4), ..., Hvd (nl), Hvd (n) esgO ", which indicates that each of the encoded frame data Pd (2), Pd (3), Pd (4), ..., Pd (nl), Pd (n) is not randomly reproducible. In each of the Hvd (l) headers, ...

Hvd (r), ..., Hvd (n), as well as the Hv header of the compressed image data Dv of the first mode shown in figure 1, the synchronous Hsd signal, the Hfdl or HfdO flag, the data Common HCD and alignment data have aligned in this order. As in the first mode, the header that includes the Hfd identification flag indicating that the compressed image DVD data is less suitable for use in random reproduction is added to the compressed image DVD data, although this is not true. show In this first method of transmitting the modification, the header of all the compressed image data is first transmitted and then the compressed frame data (encoded frame data) of the respective frames are transmitted sequentially together with the header frames of the frames. corresponding framework. When transmitting the header of all the compressed image data, or frame Hvc or Hvd headers, the synchronous Hsd signal, the identification flag Hfd or the flag Hfdl or HfdO, the common Hcd data and the data are transmitted Had of alignment, in this order. An image processing method according to the modification will be described. Figure 6 is a diagram for explaining the image processing method of the modification to which the image processing method of the first embodiment of Figure 2 has been extended. In this method of image processing of the modification, the steps 105-107 included in the processing method of the first embodiment of figure 2 are defined by step 205, step 209, step 210 and step 207, shown in figure 6. The other steps 201-204 and 208 included in the method of modification are identical to step 101-104 and 108 included in the method of the first embodiment.

This will be described in detail later. Returning to Figure 6, the coding process begins (step 201). The synchronous sequence signal Hsd is generated indicating the start of the compressed image data Dv (step 202). The synchronous Hsd signal is represented by a unique code (32 bits). Subsequently, a code of the identification flag Hfd is generated (step 203). When all the frames of the digital moving illustration are to be compressively encoded without reference to other frames, the value of the identification flag Hfd is 5 £ l ", whereas, when they are not going to be, the value is $ ¡¡ 0". A code of the common data necessary to reproduce the compressed image Dv data at the player end and an alignment data code is produced (step 204). The frame data of the respective frames of the moving artwork is compressively encoded sequentially (steps 205, 209, 210, 207). To be more detailed, the frame data is introduced (step 205). A synchronous frame signal is generated (step 209). The synchronous frame signal is different from the synchronous sequence signal in step 202 where it is represented by a unique code indicating the start of each frame. A flag indicates whether the frame data encoded or not of an objective frame is reproducible independently (step 210). In accordance with the value of the identification flag Hfd and the value of the flag indicating whether the encoded frame data is independently reproducible or not, the input frame data is compressively encoded to produce coded frame data (step 206) . The coding process in step 206 is identical to that of step 106 in the first mode. To be specific, when the value of the identification flag Hfd is 1", the entire frame data P (l) -P (n) of the moving picture image data D are encoded within the frame, whereas when the value of the Hfd flag is §0", according to the flag in relation to the possibility of independent reproduction, the frame data specified in the illustration D in motion are encoded within the frame and the other frame data they are encoded between the frames. After step 206, it is decided whether the input frame data is the last frame or not (step 207). When it is decided that the input frame data is not the last frame, the generation of the frame synchronous signal is performed again in step 209, the generation of the flag in relation to the possibility of independent reproduction in step 210 and the coding process in step 206, or otherwise, is contemplated in the coding process (step 208).

The coding process performed in this way to produce already the compressed image Dvc data for use in the random reproduction as shown in FIG. 5 (a), or the compressed image Dvd data with a high coding efficiency, the which are less suitable for use in random reproduction as shown in Figure 5 (b). This compressed image data is transmitted to the decoding end by means of the communication line or stored in the storage medium to be supplied to the decoding end. According to the modification of the first mode, in addition to the coding process performed by the image processing method of the first mode, the frame header (auxiliary header) is added to the encoded frame data (compressed frame data). of each of the frames of the compressed image data, the frame header includes the flag indicating whether the modified frame data is reproducible independently or not, and from the start code of the frame header through the code immediately before the length of the flag is set. Therefore, it is possible to quickly analyze the identification flag included in the header of the complete compressed image data and the flag included in the frame header, when analyzing the frame header added to the encoded frame data. In this modification, when the fast-forward playback process is performed on the compressed image data, according to the flag included in the frame header, it is quickly decided whether the coded frame data of each frame is reproducible independently or not. , so that a rapid reproduction process can preferably be carried out.

[MODALITY 2] Next, an image processing method according to a second embodiment of the invention will be described. In this image processing method, compressed image data which is obtained by compressively encoding digital image data corresponding to an image comprising a plurality of frames, is decoded to provide reproduced image data corresponding to the image. In this second embodiment, the premise is established that the compressed image data to be decoded is the compressed image DV data having the data structure shown in FIG. 1 (a). To be specific, the compressed image Dv data is compressed image Dva data suitable for use in random reproduction (figure 1 (b)) or compressed image DVD data which are less suitable for use in random reproduction but have a high coding efficiency (Figure 1 (c)). Figure 7 is a diagram for explaining the method of image processing according to this second embodiment, which illustrates the flow of a decoding process by the image processing method. Initially, when entering the compressed image Dv data, which have been encoded by the image processing method of the first embodiment (data specific structure of which is shown in Table 4 - Table 6) (step 301), the synchronous signal sequence in the header added to the compressed image data Dv is detected in the coding process (step 302) . This synchronous sequence signal corresponds to the data 802 in Table 4. Subsequently, according to a control signal provided by external operation, the reproduction or not of random access for early reproduction, fast rewind playback, or editing of image (step 303). The control signal is provided by external input, such as the pressing operation of the fast advance button.

When the result of the decision in step 303 is that the random access reproduction is not performed, the common data in the header (data 803-815 in Tables 4-6) are analyzed to prepare the decoding of the corresponding coded frame data for the respective frames (step 307). Subsequently, the frame data encoded from the respective frames are reproduced by a predetermined decoding method (step 310). In this second embodiment, for the frame data encoded in frame I, the quantization coefficients corresponding to each sub-block are subjected to inverse quantization and inverse DCT to reproduce image data corresponding to each of the macroblocks, and this process is performs for all the macroblocks that constitute the frame. For the frame data encoded in frame P or frame B, decoding between frames is performed with reference to data reproduced from another frame. To be specific, in the decoding between frames, the quantization coefficients (quantified DCT coefficients) of each sub-block are subjected to inverse quantization and inverse DCT to restore the difference of data corresponding to each macroblock. Subsequently, prediction data is generated for image data corresponding to an objective macroblock (macroblock to be decoded) in a frame that is going to be processed, by movement compensation from image data of a frame which has already been decoded. The predicted data and the restored difference data are added to reproduce image data of the target macroblock. Subsequently, it is decided whether or not the encoded frame data to be decoded corresponds to the last frame in the compressed image Dv data (step 311). When the result of the decision is that the encoded frame data does not correspond to the last frame, the decoding process is carried out again in step 310. When the encoded frame data corresponds to the last frame, the decoding of the frames ends. compressed image Dv data (step 312). On the other hand, when the result of the decision in step 303 is that a random access reproduction is to be performed, initially it is decided whether the compressed image data is decoded or not and if it is suitable for use in random reproduction. That is, the identification flag Hfd is extracted which has already been described for the decoding process (step 304). In this second embodiment, the data 814 of the identification flag Hfd is placed after the data 802 (synchronous sequence signal). Therefore, it can extract the Hfd flag from identification immediately after the analysis of the synchronous sequence signal. Subsequently, the value of the identification flag Hfd is checked to decide whether the compressed image input Dv data is suitable or not for use in the random access reproduction (independent reproduction) (step 305). When the value of the identification flag Hfd is §1"this Hfd flag indicates that the compressed image Dv data is suitable for use in independent reproduction, while the value of the identification flag Hfd is §0", this flag Hfd indicates that the compressed image Dv data is less suitable for use in independent reproduction. When the result of the decision in step 305 is that the value of the identification flag Hfd is §1", the common data, which follows the identification flag Hfd and which is related to the image processing of the frames respective, are analyzed (step 308) Subsequently, the encoded data corresponding to the respective frames are reproduced by decoding (step 310) The decoding process in step 310 in this case is different from the decoding process in the case in the one that does not realize the random access, only that the decoding between frames is not realized.

Subsequently, it is decided whether or not the encoded frame data to be decoded corresponds to the last frame in the compressed image Dv data (step 311). When the result of the decision is that the decoded frame data does not correspond to the last frame, the decoding process is carried out again in step 310. When the encoded frame data corresponds to the last frame, the decoding of the frame is completed. the compressed image Dv data (step 312). On the other hand, when the result of the decision in step 305 is that the value of the identification flag Hfd is S ° "that is, when the compressed image data Dv is less suitable for use in independent reproduction, it is transmitted a message describing that independent reproduction will not be performed (step 306), and the decoding of the input compressed image Dv data is completed (step 312) As described above, in the image decoding process of According to the second embodiment, since the header of the input compressed image data Dv is set up so that the identification flag Hfd is placed just after the synchronous sequence signal, it is possible to decide immediately whether the data Dv of Compressed image correspond to a moving illustration are suitable or not for use in reproduction independent, that is, whether all of the encoded frame data corresponding to the respective frames constituting the moving illustration are reproducible or not independently. Although in this second embodiment the detection of the synchronous sequence signal is performed (step 302) before the random access decision (step 303), step 302 may be carried out after step 303. In addition, although in this second mode the identification flag Hfd is placed just after the synchronous sequence signal in the header of the compressed image data Dv to be decoded, the identification flag Hfd can be placed after the N-bit data of fixed length in the header of the compressed image Dv data. In this case, when the random access is performed, the identification flag Hfd is extracted without extracting the N-bit data in step 304, and analysis of the data including the data of N is carried out in step 308. -bitio In addition, when the identification flag Hfd is placed before the data with the decision condition (fixed length coding data) or the variable length coding data, the flag indicating whether the data is reproducible independently or can not be extract immediately without having to decide the condition, and so Therefore, it is suitable for random access. Especially in the case where the identification flag Hfd is placed before the variable length code, when analyzing the data before the identification flag Hfd, it is not necessary to compare those of the input header with the data prepared in the table , so you can immediately extract the Hfd identification flag. In addition, the header structure of all the compressed image data Dv is not restricted to those described above in which the identification flag is placed just after the synchronous Hfd signal. As shown in Figure 5, a frame header including a flag can be provided in relation to the possibility of independent reproduction to each of the encoded frame data constituting the compressed image data. In this case, since there is a flag for each frame, it is only sufficient to check the coded frame data corresponding to each frame to decide whether this frame is independently reproducible or not. Furthermore, in this case, since the flag is placed just after the synchronous frame signal in the frame header, it can decide in a short time whether the frame data is reproducible independently or not. Now an image processing apparatus (image decoding apparatus) will be described for performing the Compressive decoding process by the image processing method according to the second embodiment. Figure 8 is a block diagram illustrating the image decoding apparatus according to the second embodiment. The image decoding apparatus 100b is adapted to receive compressed image data 511 which is obtained by compressively encoding digital image data corresponding to an image comprising a plurality of frames, and decoding the compressed image data to provide data of reproduced image that correspond to the image. In this second embodiment, the premise is established that the compressed image data is generated by the image coding apparatus 100a according to the first embodiment. To be specific, the image decoding apparatus 100b includes an analyzer 502 and a data decompressor 503. The analyzer 502 analyzes a header and the other data included in the compressed image data 511 to generate control information 523 or motion information 524, and analyzes the data corresponding to each frame included within the compressed image data 511 to transmit 512 compressed frame data. The decompressor 503 decompresses the data 512 of frame tablets corresponding to each frame, to generate decompressed frame data 514. In this second embodiment, the analyzer 502 is constructed so that it analyzes the identification flag without analyzing the common data constituted of fixed length codes that extend from the start of the header before the identification flag. The data decompressor 503 comprises an inverse quantizer 503a and an inverse DCT unit 503b. Inverse quantizer 503a subjects the compressed image data to inverse quantization to generate data 513 in the frequency domain, and inverse DCT unit 503b subjects the quantizer output 503a to inverse DCT to transform the data in frequency domain to data in space domain and transmit the decompressed frame data 514. The image decoding apparatus 100b further includes a prediction data generator 506 and an adder 505. The prediction data generator 506 generates prediction frame data 520 based on the output of decompressed frame data of the data decompressor 503. . The additive 505 adds the decompressed frame data corresponding to the target frame and the prediction frame data 520 corresponding to the frame data output 515 reproduced to an output or transmition terminal 510.

In the image decoding apparatus 100b, the motion information 524 (motion inspector) obtained by analysis of the analyzer 502 is supplied through a first switch 522a to the prediction data generator 506, and the prediction frame data 520. are supplied through the second switch 522b to the adder 505. In addition, the reproduced frame data 515 transmitted from the adder 505 is supplied through the third switch 522c to the prediction data generator 506. In addition, the image decoding apparatus 100b includes a controller 521 which controls the switches 522a-522c through the use of control signals 525a-525c, based on the control information 523 that is obtained by analysis of the header by the analyzer 502. The prediction data generator 506 will be described in greater detail. The prediction data generator 506 includes a frame memory 507 which stores the output 515 (reproduced data) of the adder 505 as reference image data for a frame to be processed later. The frame memory 507 transmits the data stored therein, according to the read address signal 518. In addition, the prediction data generator 506 includes a steering generator 508 and a driving unit 509. reading of prediction signal. The address generator 508 generates the read address signal 518 to the frame memory 507, based on the motion vector 517 from the analyzer 502. The prediction signal reading unit 509 obtains data in an area specified by the read address signal 518 from the frame memory 507, and transmits it to the prediction frame data 520. The operation of the image decoding apparatus 100b will be described. When the compressed image data having the format shown in Table 4 - Table 6 which have been compressively encoded by the coding apparatus 100a of the first embodiment, is transmitted to the input terminal 501, initially, the Analyzer 502 analyzes the header of the compressed image data to detect the synchronous signal of sequence and the like. Meanwhile, an external control signal (not shown) is input to the analyzer 502 indicating that random access is to be performed or not for fast forward playback, fast rewind playback or image editing, by means of a external input such as a press operation or a fast forward button. When random access is not performed, the 502 parser analyzes the common data (data 803-815 in Tables 4-6) to prepare to decode the data from coded frames of the respective frames. In this case, the switches 522a-522c are each controlled according to the type of frame coding (illustration I, illustration P or illustration B) by the control signal from the controller 521. In addition, the analyzer 502 extracts the vector 524 of motion based on the frame data encoded by each frame in the compressed image data and, in addition, transmits a quantization scale and quantization coefficients of each frame as data 512, compressed to the data decompressor 503. When the encoded frame data has been obtained by the coding process in frame, the compressed data 512 (quantization coefficients) is subjected to inverse quantization by the quantizer 503a in the data decompressor 513 to be transformed into data 513. frequency domain, which are subjected to inverse DCT by the inverse DCT unit 503b to be transformed to domain-space data (decompressed data) 514. This inverse quantization and inverse DCT is performed for each of the sub-blocks that constitute one of the macroblocks that comprise a frame. Subsequently, the decompressed data 514 corresponding to each of the macroblocks passes through the adder 505 and is transmitted as data 515 reproduced through the output terminal 510.

When the encoded frame data has been obtained by the coding process between frames, in the data decompressor 503, the compressed data (quantization coefficient) corresponding to each sub-block is subjected to inverse quantization and inverse DCT, and in this way the difference data corresponding to each macroblock is restored as decompressed data. Subsequently, the additive 505 adds the difference data to a target macroblock to the corresponding prediction data and transmits the reproduced data 515. In the prediction data generator 506, the reproduced data of a frame which has already been decoded and stored in the frame memory 507 is read from the frame memory 507 as prediction data for the target monoblock in the frame that is will decode, by movement compensation using the steering generator 508 and the prediction signal reading unit 509, as well as the prediction process of the coding apparatus. That is, in the prediction data generator 506, the address generator 508 generates the address signal 518 of the frame memory 507, based on the motion vector of the analyzer 502. Subsequently, the signal reading unit 509 The prediction data obtains data in the area specified in the direction signal as the prediction data 520.

Subsequently, the addition 505 adds the difference data of the data decompiler 503 and the prediction data and transmits the reproduced data of the macroblock or jetivo. On the other hand, when random access is performed, the identification flag in the header of loe datoe tablets to be decoded is analyzed to decide whether the compressed image data is suitable or not for use in independent reproduction. That is, the analyzer 502 extracts the flag 523 from identification and transmits it to the controller 521. In the header of the compressed image data, the data 814 that it codes correspond to the flag Hfd of identification are placed just after the data 802 that they correspond to the synchronous signal of sequence and, therefore, the identification flag Hfd can be extracted in a short time. When the identification flag Hfd is input to the controller 512, the controller 521 analyzes the identification flag Hfd to decide whether the compressed image data is suitable for use in random reproduction. When you decide that the compressed image data is suitable for use in independent reproduction (ie, when the value of the identification flag Hfd is §1"). the analyzer 502 analyzes (performs a variable length decoding of) the common data and, subsequently, the data decompressor 503 performs the reproduction to provide the reproduced data of encoded frame data that correeponden to a specified frame. In the decoding process, since the decoding between frames is not carried out, unlike the decoding process in which random access is not performed, the switches 522a-522c are kept in the OFF state by the signals of control from the controller 521. When the value of the identification flag Hfd is ijj? O "(ie, when the compressed image data is less suitable for use in independent reproduction), information is provided describing that the decoding independent should not be performed, from the controller 521 to the desired parts of the decoding apparatus, as described above, according to the image decoding apparatus of the second mode, since the analysis of the synchronous signal in the Data header compressed image Dv is followed by the analysis of the Hfd identification flag, it is possible to decide immediately if the Dv data of compressed image that correspond to a moving illustration are suitable or not for use in independent reproduction, that is, if all the encoded frame data corresponding to frames repectivoe that they constitute the iluetración in movement eon reproduciblee independently or not. When an image processing program is registered to perform the image processing according to any of the modalities described above and modifications of the programming elements (software) are recorded in a data storage medium such as a floppy disk, it is You can easily implement image processing in a separate computer system. Figures 9 (a) -9 (c) are diagrams to explain the case in which an image processing is executed according to any of the modalities and modifications described above, by a computer system, using a flexible disk which contains the image processing program. Figure 9 (a) shows a front view of a flexible disk FD, a cross-sectional view thereof, and a flexible disk body D. Figure 9 (b) shows an example of a physical format of the floppy disk body D. The flexible disk FD has the configuration in which the flexible disk container FC contains the flexible disk body D. On the surface of the flexible disk body D, a plurality of tracks Tr is concentrically formed from the outer circumference of the disk towards the inner circumference. Each track Tr is divided into 16 sectors (Se) in the angular direction. Therefore, in the flexible disk FD having the aforementioned program, the program data is recorded in the sectors assigned in the body D of the floppy disk. Figure 9 (c) illustrates the structure for recording the program on the flexible disk FD and performing the image processing by programming elements (software) using the program stored on the flexible disk FD. To be specific, when the program is recorded on the FD floppy disk, the program data is written to the FD floppy disk from the Cs computer system through the FDD floppy disk drive. When the image coding apparatus described above or the image decoding apparatus is constructed in the computer system Cs by the program recorded on the flexible disk FD, the program is read from the flexible die FD by the flexible disk unit FDD and then it is loaded into the computer system Cs. Although a flexible disk is used in the above description as the data storage medium, an optical disk can be used. Furthermore, in this case, the coding or decoding by the programming elements (software) can be done in a manner similar to the case in which the floppy disk is used. In addition, the data storage medium is not reetrinated to these discs. It can be used any means to the extent that the program can contain, for example, an IC card, a ROM cassette, etc. Furthermore, in the case of using this data storage means, an image processing by programming elements (software) can be performed in a manner similar to the case of the use of a flexible disk. Further, when an encoded image signal stored in a data storage medium such as a floppy disk has a data structure according to any of the embodiments and modifications of the invention, the encoded image signal of the floppy disk can be decoded. for image display, with the same ones as those described for the decoding process of the first or second modality.

INDUSTRIAL AVAILABILITY According to the image transmission method, the image processing method, the image processing apparatus and the data storage medium of the present invention, compressed image data can be selected and reproduced randomly from an arbitrary frame of an illustration in motion without waiting time. Therefore, this is very useful in image coding processes and decoding processes of image in the seventh which tranemite or stores an image signal, and in particular, is suitable for use in a random image processing for arbitrary frames such as fast forward playback, fast rewind playback or motion picture data editing compressed , according to MPEG4. It is noted that in relation to this date, the best method known to the applicant to carry out the aforementioned invention, is that which is clear from the present description of the invention.

Claims (22)

CLAIMS Having described the invention as above, property is claimed as contained in the following claims

1. An image transmission method for transmitting compressed image data obtained by compressively encoding digital image data corresponding to an image comprising a plurality of frames, the method is characterized in that it comprises the steps of: transmitting a header including common data to respective frames which are included in the compressed image data; and sequentially transmitting compressed frame data of the respective frames which are included in the compressed image data, wherein the header includes an identification flag which indicates whether the compressed image data is suitable or not for use in a process of random reproduction, which randomly selects data from arbitrary frame compressed frames to reproduce frame data, and from code data at the beginning of the header to encode data immediately before the identification flag is set to length.

2. The method of image transmission, according to claim 1, characterized in that the header comprises a synchronous signal indicating a header of the compressed image data, fixed-length code data and variable-length code data as the common data to the respective frames, and the identification flag, and in which in the header transmission step, the identification flag is transmitted after the synchronous signal and before the variable length coding data.

3. An image processing method for compressively encoding digital image data corresponding to an image comprising a plurality of frames, the method comprising the steps of: generating a header which includes data common to the respective data; and compressively encoding frame data corresponding to respective frames to produce compressed frame data, wherein the header is generated in such a way that it includes an identification flag that indicates whether the compressed image data is suitable or not for use in a Random playback process which randomly selects compressed frame data from arbitrary frames to reproduce frame data, and from code data to principle of the header to encode data immediately before the identification flag is set to length.

The method of image processing, according to claim 3, characterized in that: the step of compressively coding frame data is performed after the step of generating the header and in which the step of generating the header comprises generating a synchronous signal indicating a header of the compressed image data, the identification flag, and the data common to the respective frames, in this order.

The method of image processing, according to claim 3, characterized in that: the step of compressively coding frame data is performed after the step of generating the header, and in which the step of generating the header comprises generating a synchronous signal indicating a header of the compressed image data, fixed-length code data such as the data common to the respective frames, the identification flag, and variable-length code data as the data common to the respective frames, in this order.

6. The method of image processing, according to claim 3, characterized in that: the step of compressively coding frame data comprises: a first step of compressive coding to compressively encode frame data corresponding to a frame to be processed without reference to frame data corresponding to another frame, to produce first frame data tablets; and a second step of compressive coding to compressively encode frame data corresponding to a frame to be processed with reference to the frame data corresponding to another frame, to produce second data of compressed frames, wherein the flag of identification it is included in the compressed image data comprised of the first compressed frame data indicating that the compressed image data is suitable for use in the random reproduction process, and the identification flag included in the compressed image data is comprised of first compressed image data and second compressed frame data indicating that the compressed image data is less suitable for use in the random reproduction process.

The method of image processing, according to claim 3, characterized in that it also comprises the step of: generate an auxiliary header that includes data common to the respective frames and individual data for a specified frame, in which the auxiliary header is generated such that the auxiliary header is added before the frame data compressed from the specified frame when the step of compressively coding frame data is performed after the step of generating in header, and in the step of generating the auxiliary header, the auxiliary header is generated in such a way that it includes a flag indicating whether the frame data compressed from the frame specified are independently reproducible or not without reference to the frame data of another frame, and from code data to the beginning of the auxiliary header to encode data immediately before the flag is set to length.

8. An image processing method for decoding compressed image data obtained by compressively encoding digital image data corresponding to an image comprising a plurality of frames that provide reproduced frame data corresponding to the image, the method is characterized in that it includes the stages of: analyzing a header that includes data common to the repective data which are included in the compressed image data, - and decoding data from compressed frames obtained by compressively coding frame data from respective frames and included in the compressed image data, to provide reproduced frame data, in which the step of analyzing the header comprises analyzing the length code data fixed from the code data at the beginning of the header to encode data immediately before an identification flag is included in the header and indicate whether the compressed image data is suitable or not for use in a random reproduction process which randomly select compressed frame data from arbitrary data to provide reproduced frame data, and then analyze the identification flag.

The method of image processing, according to claim 8, characterized in that: the step of decoding compressed frame data is performed after the step of analyzing the header, and in which the step of analyzing the header comprises analyzing a synchronous signal indicating a header of the compressed image data, the identification flag, and the data common to the respective frames, in that order.

10. The method of image processing, according to claim 8, characterized in that: the step of decoding compressed frame data is performed after the step of analyzing the header, and wherein, the step of analyzing the header comprises analyzing a signal synchronous indicating a header of the compressed image data, fixed-length code data as the data common to the respective frames, the identification flag and the variable length code data as the data common to the respective frames, in this order.

The method of image processing, according to claim 8, characterized in that: the step of analyzing the header and the step of decoding the compressed frame data are performed for the first compressed image data constituting the first data of frame compresses obtained by compressively encoding frame data to a frame to be processed without reference to the frame data of another frame, and to perform for the second compressed image data consisting of the first frame data compressed and the second compressed frame data obtained by compressively encoding frame data from a frame that goes to be processed with reference to the frame data of another frame, and the random reproduction process is performed only for the first compressed image data, according to the identification flag.

The method of image processing, according to claim 8, characterized in that it further comprises the step of: analyzing the auxiliary header added to the compressed frame data of a specified frame and including data common to the respective frames and data individual for the specified frame, in which the auxiliary header is analyzed by the specified frame when performing the stage of decoding compressed frame data after the stage of analyzing the header, and the step of analyzing the auxiliary header comprises analyzing the data of fixed-length code from the code data at the start of the auxiliary header to encode data immediately before the flag is included in the auxiliary header and indicating whether the compressed frame data of the specified frame is independently reproducible or not reference to the frame data of another frame, and then analyze the flag.

13. An image processing apparatus for compressively encoding digital image data corresponding to an image comprising a plurality of frames for producing compressed image data, the apparatus is characterized in that it comprises: a prediction data generator for generating frame data; prediction for objective frame data corresponding to a frame to be processed based on the target frame data, - a calculation means for transmitting either the difference frame data as a difference value between the frame data target and frame data prediction of target frame data, according to a control signal; a data compressor for compressing the data output from the computing means to produce compressed data; a variable length encoder for performing variable length encoding of the output of compressed data from the data compressor and transmitting the compressed frame data of each frame, and a control means for generating a header that includes data common to the frames respective based on the digital image data and to control the calculation means according to an identification flag indicating whether the Compressed image data is suitable or not for a random reproduction process which randomly selects compressed frame data from arbitrary frames to reproduce data frames, wherein the variable length encoder transmits the header including the identification flag, in which from the code data at the beginning of the header to the code data immediately before the identification flag are set as to length.

The image processing apparatus, according to claim 13, characterized in that the variable length encoder transmits the header before transmitting the compressed frame data of the respective frames in such a way that the synchronous signal indicating a signal is transmitted. header of the compressed image data, the identification flag and the data common to the respective data, in this order.

The image processing apparatus, according to claim 13, characterized in that the variable length encoder transmits the header before transmitting the compressed frame data of the respective frames in such a way that a synchronous signal indicating a header of the compressed image data, fixed length coding data such as common data for the reepectivoe marcoe, the flag of identification and the data of variable length code as the data common to the respective frames, in this order.

16. The image processing apparatus according to claim 13, characterized in that the control means controls the calculation means so that all the frames of the image are subjected to a first compressive coding process in which the medium of calculation transmits the target frame data, the data compressor compresses the frame data of a frame to be processed without reference to frame data of another frame, and the variable length encoder transmits the first compressed frame data, when the identification flag indicates that the compressed image data is suitable for use in a random reproduction process, and wherein the control means controls the computing means so that the frames specified in the image are subjected to a second compressive coding process in which the calculation medium transmits the difference frame data, the data compressor compresses the frame data e a frame to be processed with reference to the frame data of another frame, and the variable length encoder transmits compressed second frame data, and the frames other than the specified frames are subjected to the first compressive coding process, when the identification flag indicates that the image data compressed eon menoe suitable for use in the process of random reproduction.

17. An image processing apparatus for decoding compressed image data obtained by compressively encoding digital image data corresponding to a frame comprising a plurality of frames to provide a reproduced image corresponding to the image, the apparatus is characterized in that: an analyzer for analyzing a header included in the compressed image data to generate header information, and analyzing data corresponding to each frame included in the compressed image data and transmitting the compressed image data, - a data decompressor for decompressing the data compressed frame data to produce uncompressed frame data, - a calculation means for transmitting either frame data that is obtained by adding the decompressed frame data and frame prediction data or the frame data decompressed as data of frames reproduced according to a control signal, - a g prediction data eager to generate prediction framework data for a framework that is to be processed from objective framework data that is going to decompress, which correspond to the frame to be processed, and a control means for controlling the calculation means according to an identification flag included in the header information and indicating whether the compressed image data is suitable or not. for use in a random reproduction process which randomly selects data from compressed frames of arbitrary frames to reproduce frame data, in which the analyzer analyzes the header in such a way that it analyzes the identification flag without analyzing common data. Fixed-length code from code data at the beginning of the header to encode data immediately before the identification flag is required.

18. The image processing apparatus according to claim 17, characterized in that the parsing analyzes the header in such a way that it analyzes a synchronous signal indicating a header of the compressed image data, the identification flag and the common data. to the respective data, in this order, according to the order in which these data are entered into the analyzer.

19. The image processing apparatus, according to claim 17, characterized in that: the analyzer analyzes the header in a manner that analyzes a synchronous signal indicating a header of the compressed image data, fixed-length code data such as data common to the respective frames, the identification flag and the variable-length code data as the data common to the respective frames, in this order, according to the order in which they are introduced to the analyzer.

20. The image processing apparatus according to claim 17, characterized in that the control means controls the calculation means so that the entire frame of the image is subjected to a first decompressive decoding process without reference to another frame. in which the decompressed frame data corresponding to the frame to be processed are transmitted from the calculation medium as reproduced frame data, when the identification flag indicates that the compressed image data is suitable for use in the processing process. random reproduction, and wherein the control means controls the calculation means so that the specified frames of the image are subjected to a second decompressive decoding process with reference to another frame in which a data addition value is transmitted of the unzipped frame or the frame to be processed and the frame data reproduced from another frame, to starting from the calculation medium as frame data reproduced from the frame to be processed and frames different from the specified frames that are submitted to the first uncompromising decoding process, when the identification flag indicates that the image data compressed are less suitable for use in the random reproduction process.

21. A data storage medium for storing an image processing program for compressively encoding digital image data corresponding to an image comprising a plurality of frames, the image processing program allows the computer to perform a coding process compressor for the digital image data according to an image processing method of claim 3.

22. The data storage means for storing an image processing program for uncompressed decoding of compressed image data obtained by compressively encoding data. of digital image corresponding to an image comprising a plurality of frames, the image processing program allows a computer to perform a decoding process for the compressed image data according to an image processing method of claim 8.