JPH08172628A - Reduction method of quantized noise caused when converted and coded image data are decoded and decoder for image data subjected to conversion coding - Google Patents

Abstract

PURPOSE: To reduce quantized noise caused when image data subjected to conversion coding are decoded by applying low pass filter processing to image data of a field by the interlace scanning. CONSTITUTION: When image data decoded by a variable length decoding section 9 and black quantization information in added information are given to an inverse quantization section 10, a DCT conversion coefficient obtained through inverse quantization is fed to an inverse orthogonal transformation section 11. The transformation section 11 applies inverse DCT for each unit block to inversely transform image data on a frequency region into image data of a time base area and the data are fed to an adder section 12. Then the image data are stored in an image memory 14 according to a coding type representing the difference from in-frame coding and inter-frame coding. Then image data in the memory 14 in a decoder integrated circuit 3 are outputted to an output terminal 2 via a low pass filter 7. Thus block distortion or quantization noise caused when an image by the interlace scanning system is decoded is reduced.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for reducing quantization noise generated when decoding transform-encoded image data, and decoding the transform-encoded image data by applying the above method. Regarding the chemical conversion device.

[0002]

2. Description of the Related Art In the case of transmitting, recording and reproducing image signals, audio signals, and other various signals as digital signals,
Information compression / expansion technology is used. That is, for example, when digitizing an image signal or an audio signal, linear quantization (equal quantization) in which each sample value is replaced with one of the representative values of the signal level that is equally divided is transmitted. ， The amount of information of the signal to be recorded and reproduced is
This is because there will be very many. Therefore, in the technical field of broadcast communication and the technical field of recording / reproduction, for example, human vision and hearing are sensitive to a portion where the signal changes little, and there is some error in a portion where the signal changes drastically. In addition to utilizing the human visual and auditory properties, which are difficult to detect, to reduce the amount of information for each sample, many information compression techniques are applied to transmit, record, and reproduce. It is well known that a high-efficiency compression technique (high-efficiency information encoding technique) for various kinds of targeted information has been put into practical use.

Now, the amount of information per hour in a moving image having a quality similar to that of a reproduced image displayed using a reproduced signal from a VHS (registered trademark) VTR currently in practical use is about 109 G bits. In addition, the amount of information per hour in a moving image of a quality similar to the received image of the current standard color television system in Japan is about 360 G bits, but it has a large amount of information as described above. Practical studies have also been actively conducted on a high-efficiency compression method of image information required for transmitting and recording / reproducing image information using a current transmission line or recording medium that has been put to practical use. ..

By the way, in the high-efficiency compression method of image information, which is currently proposed as a practical high-efficiency compression method of image information, the correlation between the adjacent pixels in the natural image is high. Of the amount of information performed by using the correlation function (compression of the amount of information performed by using the spatial correlation function) and the amount of information performed by using the correlation function between screens (frames) arranged on the time axis. The amount of information is compressed by combining three different types of compression means: compression (compression of information amount using temporal correlation), compression of information amount due to deviation of code appearance probability, and compression of information amount. Efficient coding is achieved. As a means for compressing the information amount of an image using the above-mentioned intra-frame (intra-frame) correlation, many methods have been proposed in the past, but in recent years, K-L ( Karonen-Reeve) transform, discrete cosine transform (DCT), discrete Fourier transform,
Orthogonal transformations such as Walsh-Hadamard transformations have been adopted as typical examples.

[0005] For example, MPEG (Moving Picture C) established under ISO (International Organization for Standardization)
The high-efficiency coding method of image information (sometimes referred to as MPEG1 method or MPEG2 method) proposed as a result of international standardization work by the Odging Expert Group includes intra-frame coding and inter-frame coding. High-efficiency coding of moving image information is performed in combination with motion-compensated prediction and inter-frame prediction. As the orthogonal transformation, two-dimensional discrete cosine transformation (2
Dimensional DCT) is adopted. In the orthogonal transform, a predetermined block size (N × M pixels ← horizontal N pixels × vertical M line block size) is set for the image signal of each screen to be subjected to high efficiency encoding. An image divided for each “unit block” (in the MPEG1 system and the MPEG2 system, a block having a block size of 8 × 8 pixels ← horizontal 8 pixels × vertical 8 lines is a “unit block”) Performed on signals.

[0006] (N × M) orthogonal transform coefficients (8 × 8 = 6 in the above-mentioned MPEG1 system and MPEG2 system) obtained by performing orthogonal transform for each block of the above unit.
The four DCT transform coefficients) are areas of a predetermined size including at least one of the blocks of the above-mentioned units (areas called “macroblock” in the above-mentioned MPEG1 system and MPEG2 system). , That is,
In the MPEG1 method and the MPEG2 method, an area having a block size of 16 × 16 pixels ← horizontal 16 pixels × vertical 16 lines for the luminance signal Y and two color difference signals Cr and C
(A region consisting of 8 × 8 pixels ← horizontal 8 pixels × vertical 8 line block size region for each b)
It is quantized by the "block quantization width value" set for each. For example, in the MPEG1 system and the MPEG2 system, the above-mentioned “block quantization width value” is represented as [{macroblock quantization characteristic value (or macroblock quantization scale) QS} × quantization matrix].

The orthogonal transform coefficient (for example, DCT coefficient) quantized by the block quantization width value is separated into its DC component (DC component) and AC component (AC component). The DC component of the orthogonal transform coefficient (for example, DCT coefficient) is differentially encoded, and the AC component of the orthogonal transform coefficient (for example, DCT coefficient) is entropy-encoded after zigzag scanning (information due to deviation of the appearance probability of the code). Quantity compression ... Variable length coding such as Huffman method). The image data converted and encoded as described above is output as a bit stream (bit string). Next, the decoding operation for the image data that has been transform-encoded as described above is performed in the reverse operation of the encoding operation described above to obtain the output image. When quantization is performed, the presence of a quantization error that cannot be avoided causes quantization noise in the output image. Then, when the complexity of the image to be encoded is large with respect to the transmission rate, the above-mentioned quantization noise greatly deteriorates the image quality.

Generally, among the quantization errors that cause the above-mentioned quantization noise, the quantization error of the low-frequency component is the output image distortion in a state where there is no correlation between unit blocks, so-called block distortion. Among the quantization errors that occur inside and cause the quantization noise, the quantization error of the high frequency component causes ringing-like output image distortion, so-called mosquito noise, around the edges. By the way, the quantization noise generated in the image as described above is particularly noticeable in the flat part of the image, and small noise is added to a place where there is a large change in the video signal level from the low range to the high range. In the case of such a quantization noise waveform, it is difficult to detect noise because the difference in sensitivity in visual characteristics is small.

However, when a large change in the video signal level exists only in the low frequency range and the small noise is added in the high frequency range, the noise is easily detected. Naturally, if a large amount of noise is added,
It goes without saying that fatalness is detected as coding deterioration regardless of whether it is in the low band or the high band. Then, as one of means for solving the problem of deterioration of image quality due to quantization noise as described above, for example, Japanese Patent Laid-Open No. 4-3720
As disclosed in Japanese Laid-Open Patent Publication No. 74-74, a solution is proposed to reduce the quantization noise by applying a post filter to the decoded image.

[0010]

DISCLOSURE OF THE INVENTION Problems to be Solved by the Invention
According to the solving means disclosed in Japanese Patent Publication No. 2074, image data to be transform-encoded is divided into blocks of a unit having a predetermined block size, and at least the blocks of the unit described above are divided. An area of a predetermined size including one (MPEG1 system, MPEG
In the two methods, the orthogonal transformation performed on the image data in each of the regions described above is performed for each of the regions referred to by the term "macroblock") on the image data for each block of the unit described above. When it is used for post-processing of the decoded image data from the conversion-coded image data in the case of (for example, the MPEG1 system), it is possible to achieve the intended purpose.

However, the image data to be transformed and encoded is divided into blocks each having a unit having a predetermined block size, and is defined in advance including at least one block of the unit. Area of large size (in MPEG1 system, MPEG2 system,
For each area referred to by the term "macroblock"), the orthogonal transformation performed on the image data in each of the areas described above is performed by the first orthogonal transformation performed on the image data for each unit block described above. Not only the mode, but the image data of a new unit block having the above-mentioned predetermined block size obtained by combining the image data of the odd field of the interlaced scanning in the image data in the above-mentioned area, and the image data in the above-mentioned area Any one of a second orthogonal transform mode in which orthogonal transform is performed on image data of a new unit block having a predetermined block size obtained by combining image data of even fields of interlaced scanning in image data. The transform coding is performed according to the orthogonal transform mode (for example, in the case of the MPEG2 system). And post-processing for the decoded image data from the image data, JP-4-37207 mentioned above
When the solution disclosed in Japanese Patent Publication No. 4 is used, a reproduced image may be unnatural in the movement in the time direction. That is, since there is no correlation between the image data of the two fields of the interlaced scan in the case of an image with a large motion, when such image data is subjected to the low-pass filter processing, the image data of another field is mixed and the time resolution is increased. The amount of the reproduced image is reduced, and the reproduced image has an unnatural movement in the time direction. Therefore, a solution that does not cause the above problems has been sought.

[0012]

DISCLOSURE OF THE INVENTION According to the present invention, an object of transform coding by one or both of a frame-structured image in which a frame is a unit of encoding and a field-structured image in which a field is a unit of encoding. The image data is divided into blocks of a unit having a predetermined block size, and the area of a predetermined size including at least one of the blocks of the units described above is described above. The orthogonal transformation performed on the image data in each area is the same as the first orthogonal transformation mode performed on the image data for each block of the unit described above for the image of the frame structure and the image data in the area described above. A block of a new unit having the above-mentioned predetermined block size obtained by combining the image data of the odd field of the interlaced scanning Orthogonal transformation is performed on the image data and the image data of the new unit block having the above-mentioned predetermined block size obtained by combining the image data of the even field of the interlaced scanning in the image data in the area. In contrast to the orthogonal transform mode of any one of the two orthogonal transform modes, the image of the field structure is adapted to be performed in the second orthogonal transform mode described above, and the orthogonal transform described above is performed. Quantized by using the block quantization width value set individually for each of the areas, and the image data obtained by decoding the image data that has been transform-coded by the predetermined coding. Decoded image data of a region where orthogonal transformation has been performed in the first orthogonal transformation mode described above to reduce noise. Regarding both the low-pass filter processing that should be performed in the vertical direction of the image and the low-pass filter processing that should be performed in the vertical direction of the image for the decoded image data of the area that has been orthogonally transformed in the second orthogonal transformation mode described above. As the low-pass filter processing of, the method of reducing the quantization noise generated at the time of decoding the transform-encoded image data, in which the low-pass filter processing is applied to the image data of the field of the image by the interlaced scanning method, and the above-mentioned In order to reduce the quantization noise in the image data obtained by decoding the image data that has been transformed and encoded by the first conversion, the first orthogonal method is applied to the low-pass filter processing performed on the decoded image data in the vertical direction of the image. With the unit block that was orthogonally transformed by the transformation mode For the corresponding decoded image data, the decoded image data selected for low-pass filtering in the vertical direction of the image is not present across the boundary of the unit block On the other hand, the low-pass filter processing to be performed in the vertical direction of the image is performed using the decoded pixel data that is continuous in the vertical direction in the unit block, and the orthogonal transformation is performed in the second orthogonal transformation mode. For the decoded image data corresponding to the unit block, the decoded image data selected for the low-pass filter processing in the vertical direction of the image does not exist across the unit block boundary. Low-pass filtering to be applied in the vertical direction of image data , Every other decoded pixel data in the vertical direction in the unit block is used, and the decoded image data selected for the low-pass filter processing in the vertical direction of the image sandwiches the boundary of the unit block. In the state of existing in the blocks of the two adjacent units, both of the blocks of the two units adjacent to each other across the boundary are the blocks of the units for which the orthogonal transform has been performed in the first orthogonal transform mode. In the case of a block of a unit of corresponding decoded image data, the low-pass filter processing to be performed on the decoded image data in the vertical direction of the image is decoded continuously in the vertical direction of each unit block. Performed with pixel data and also selected for low pass filtering in the vertical direction of the image. In a state where the decoded image data exists in two unit blocks that are adjacent to each other across the boundary of the unit block, at least one of the two unit blocks that are adjacent to each other across the boundary is the second block. In the case of the block of the unit of the decoded image data corresponding to the block of the unit that has been orthogonally transformed by the orthogonal transformation mode of, the decoded image data should be applied in the vertical direction of the image. A low-pass filtering process is performed using every other vertical decoded pixel data in a block of each unit, and a method of reducing quantization noise generated at the time of decoding transform-coded image data, and Image data that has been transformed and coded by prescribed coding by implementing the method for reducing the quantization noise that occurs when decoding the image data Decoding of transform-coded image data including at least a buffer memory, a variable-length decoding unit, an inverse quantization unit, an inverse orthogonal transform unit, and an image memory for decoding Orthogonal transform performed on the image data in each of the regions in the encoding device from the transform-encoded image data, for each region of a predetermined size including at least one of the blocks of the units described above. Is a predetermined block size obtained by combining the first orthogonal transformation mode performed on the image data of each block of the unit described above and the image data of the odd field of the interlaced scanning in the image data in the region. A combination of the image data of a new unit block having the above and the image data of the even field of the interlaced scanning in the image data in the above area Orthogonal transform mode information indicating whether the obtained image data of a new unit block having a predetermined block size is subjected to any of the orthogonal transform modes of the second orthogonal transform mode in which the orthogonal transform is performed. And image structure information / orthogonal transform mode information detecting means for detecting image structure information, a low-pass filter for passing the decoded image data read from the image memory, and the image structure information / information and orthogonal transform mode information detecting means. Depending on the detection power, the low-pass filter always performs low-pass filter processing on the image data of the field of the image by the interlaced scanning method, or the image structure information and the orthogonal transformation mode information detected by the variable length decoding unit. Pass the decoded image data read from the image memory based on A low-pass filter capable of performing a low-pass filtering process on image data of an image field by an interlaced scanning method, and a decoding device for transform-encoded image data, and transform-encoding by predetermined encoding In order to decode the processed image data, at least a buffer memory, a variable length decoding unit, an inverse quantizing unit, an inverse orthogonal transform unit, and an image memory are transformed and encoded. In the image data decoding device, the image data in each of the above-mentioned areas is executed for each area of a predetermined size including at least one block of the above unit from the converted and encoded image data. The orthogonal transformation is performed on the first orthogonal transformation mode performed on the image data of each block of the above-mentioned unit and within the above-mentioned area. Image data of a block of a new unit having a predetermined block size obtained by combining the image data of the odd field of the interlaced scanning in the image data of the even field of the interlaced scanning in the image data of the area The image data of the block of the new unit having the predetermined block size obtained by combining the image data and the second orthogonal transform mode in which the orthogonal transform is performed are performed according to any orthogonal transform mode. Image structure information and orthogonal transform mode information detecting means for detecting the orthogonal transform mode information and the image structure information, a low pass filter for passing the decoded image data read from the image memory, the image structure information and orthogonal transform described above. Image structure information and orthogonal transform mode information detected by the mode information detecting means, or Based on the image structure information and the orthogonal transformation mode information detected by the variable length decoding unit, the decoded image data corresponding to the unit block that has been orthogonally transformed in the first orthogonal transformation mode is the vertical direction of the image. The decoded image data selected for the direction low-pass filter processing is applied to the decoded image data in the vertical direction of the image in a state where the decoded image data does not exist across the unit block boundary. The low-pass filter processing to be performed should be performed using decoded pixel data that is continuous in the vertical direction in the unit block, and is also decoded corresponding to the unit block for which the orthogonal transform has been performed in the second orthogonal transform mode. For image data, the decoded image data selected for low-pass filtering in the vertical direction of the image. However, the low-pass filter processing to be performed on the decoded image data in the vertical direction of the image in the state where the block does not exist across the boundary of the unit block is 1 in the vertical direction in the unit block. The decoded image data selected for low-pass filter processing in the vertical direction of the image, as is performed using every other decoded pixel data, has two unit blocks adjacent to each other across a unit block boundary. In the state in which the two blocks of two units adjacent to each other across the boundary are present, the two blocks of the decoded image data corresponding to the block of the unit for which the orthogonal transformation has been performed in the first orthogonal transformation mode are performed. In the case of a unit block, the low pass buffer to be applied to the decoded image data in the vertical direction of the image. The filtered image data selected for the low-pass filter processing in the vertical direction of the image is the block of the unit so that the filtering process is performed using the decoded pixel data that is continuous in the vertical direction in each unit block. In the state of existing in the blocks of two units adjacent to each other across the boundary of, at least one of the blocks of the two units adjacent to each other across the boundary is
In the case of the block of the unit of the decoded image data corresponding to the block of the unit which has been orthogonally transformed in the second orthogonal transformation mode, the decoded image data is applied in the vertical direction of the image. Decoding transform-encoded image data, comprising means for operating the low-pass filter so that the low-pass filtering process to be performed is performed using every other vertical decoded pixel data in each unit block. Provide a device.

[0013]

With respect to each one of the frame-structured images whose frame is the unit of encoding and the field-structured image whose field is the unit of encoding, one or both of which are the targets of high-efficiency encoding. For the image signal of the screen, for each “unit block” having a predetermined block size (N × M pixels ← horizontal N pixels × vertical M line block size),
A “block quantization width value” in which (N × M) orthogonal transform coefficients obtained by orthogonal transform are set for each area of a predetermined size including at least one block of the unit described above. A bit stream composed of transform-encoded image data that has been quantized by the following and then subjected to predetermined encoding, and additional information required at the time of decoding the transform-encoded image data is a buffer memory. Memorized in. In the variable length decoding unit to which the bit stream read out from the buffer memory is supplied, the entropy coded (variable length coded) image data and the addition required when decoding the transform coded image data are added. Information {for example, block quantization width information, motion vector, prediction mode information, image structure (picture structure) information, orthogonal transformation mode information, etc.} is decoded.

Decoded image data and block quantization width information in the decoded additional information are given to the dequantization section, so that the dequantization operation performed on the image data by the dequantization section causes An orthogonal transform coefficient is supplied from the inverse quantization unit to the inverse orthogonal transform unit. In the inverse orthogonal transform unit,
The image data in the frequency domain is inversely transformed into the image data in the time domain by performing a two-dimensional inverse orthogonal transformation for each unit block. The image data in the time domain which is output from the inverse orthogonal transform unit is added to the image data in the motion compensated state by the motion compensation unit according to the coding type indicating the difference between intraframe coding and interframe coding. Or,
The output image data is stored in the image memory with or without addition.

Image structure (picture structure) information (information indicating whether the image structure is a frame structure or a field structure) detected from the additional information contained in the bit stream and the orthogonal transform mode information. The pixel address information is also given to the control signal generating section to which is supplied. In the control signal generator, the low-pass filter processing to be performed on the image data in the vertical direction of the image is performed by the low-pass filter (low-pass filter) according to the operation mode set in the device. An address signal required to output data to the output terminal 2 is supplied to the image memory. That is, in one of the operation modes set in the apparatus, the low-pass filter processing to be performed in the vertical direction of the image on the decoded image data of the region that has been orthogonally transformed by the first orthogonal transformation mode, For the decoded image data in the area that has been subjected to the orthogonal transformation in the second orthogonal transformation mode, both the low-pass filtering processing to be performed in the vertical direction of the image and the low-pass filtering processing are performed.
As illustrated in FIG. 3, it may be performed as a low-pass filter process for image data of a field of an image by the interlaced scanning method.

Next, referring to FIG. 4 which shows an example in which the first area and the third area have a frame structure and the second area has a field structure, another area set in the apparatus is described. In the case of the operation mode, the decoded image data corresponding to the block of the unit for which the orthogonal transformation has been performed in the first orthogonal transformation mode is the decoding selected for the low-pass filter processing in the vertical direction of the image. The low-pass filtering process to be performed on the decoded image data in the vertical direction of the image in a state where the image data does not exist across the boundary of the unit block is the first Vertical 1 in area
As illustrated in the range of 0 to 15e, the decoding pixel data that is continuous in the vertical direction in the unit block is used.

Regarding the decoded image data corresponding to the block of the unit which has been orthogonally transformed by the second orthogonal transformation mode, the decoded image data selected for the low-pass filter processing in the vertical direction of the image is The low-pass filtering process to be performed on the decoded image data in the vertical direction of the image in the state where the block does not exist across the boundary of the unit block is the vertical direction in the second region of FIG. 1e to 16e of the above, the image data is decoded using every other vertical decoded pixel data in the unit block, and further selected for the low-pass filtering in the vertical direction of the image. In the state where the decoded image data is present in two unit blocks that are adjacent to each other across the unit block boundary, In the case where both of the two unit blocks adjacent to each other across the boundary are the unit blocks of the decoded image data corresponding to the unit block for which the orthogonal transform is performed in the first orthogonal transform mode, Four
As illustrated in the range of 7o to 10e in the vertical direction in the first region of the above, the low-pass filtering process to be performed on the decoded image data in the vertical direction of the image is performed for each unit. This is performed using decoded pixel data that is continuous in the vertical direction in the block.

Furthermore, in the state where the decoded image data selected for the low-pass filter processing in the vertical direction of the image is present in two unit blocks adjacent to each other across the boundary of the unit block, In the case where at least one of the two unit blocks adjacent to each other across the boundary is a unit block of decoded image data corresponding to the unit block for which the orthogonal transform was performed in the second orthogonal transform mode, 4 ranges from 14e in the vertical direction in the first area of FIG. 4 to 5o in the second area of FIG. 4, and 12e in the vertical direction in the second area of FIG. 4 to 4e of the third area in FIG. The low pass filter processing to be performed on the decoded image data in the vertical direction of the image is performed on each unit block as illustrated in the range of Low-pass filter processing on the image data is performed by using the decoded pixel data of every other vertical that.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A method of reducing quantization noise generated when transform-encoded image data is decoded and a transform-encoded image data decoding apparatus according to the present invention will be described below with reference to the accompanying drawings. The contents will be described in detail. 1 and 2 are block diagrams showing a configuration example of a decoding apparatus for image data that has been transform-coded by applying a method for reducing quantization noise that occurs when decoding transform-coded image data according to the present invention. is there. In the image data decoding apparatus of the present invention shown in FIGS. 1 and 2, reference numeral 1 denotes an input terminal of a coded bit stream (coded data string) to be decoded,
Further, in each drawing, a portion 3 surrounded by a one-dot chain line frame is a constituent portion integrated into a decoder integrated circuit. In the embodiment of the decoding device shown in each of FIG. 1 and FIG. 2, at least the buffer memory 8 is provided as the component surrounded by the chain line frame 3.
A variable-length decoding unit 9, an inverse quantization unit 10, an inverse orthogonal transformation unit 11, an addition unit 12, a motion compensation unit 13, and an image memory 14 are used as an integrated circuit commercial product. can do.

The coded bit stream supplied to the above-mentioned input terminal 1 is compressed by the orthogonal transformation using the intra-frame (in-frame) correlation coefficient to compress the information amount of the image (using the spatial correlation coefficient). (Compression of the amount of information) and compression of the amount of information performed by using correlations between screens (frames) arranged on the time axis (compression of the amount of information performed by using correlations of time), and Image data (for example, MPEG1 system, MPEG) which is highly efficient conversion coded by combining three different types of compression means, that is, compression of information amount due to bias of appearance probability.
Image data according to the two systems). In addition,
In the following description of this specification, it is described that the image data to be decoded is the image data according to the MPEG1 system and the MPEG2 system.

High-efficiency encoding of moving image information in the MPEG1 system is performed on image data of an image having a frame structure, and high-efficiency encoding of moving image information in the MPEG2 system is a unit of encoding a frame. The image data of one of the image data of the frame structure image and the image of the field structure of which the field is a unit of encoding are MPE.
High-efficiency coding of moving image information in the G1 and MPEG2 systems is a combination of intraframe coding by two-dimensional discrete cosine transform (two-dimensional DCT) and interframe coding,
It is performed with motion compensation prediction and inter-frame prediction. Then, the image signal of the screen for each one which is the target of high efficiency encoding is 8 × 8 pixels (8 pixels horizontally × 8 lines vertically).
It is divided for each "unit block" of the block size of
DCT is performed for each block of each unit. And 64 DCs for each block of each unit described above.
The T transform coefficient is quantized by the "block quantization width value".
In the MPEG1 system and the MPEG2 system, the "block quantization width value" is a "macro block" of an area of a predetermined size including one of the blocks of the unit described above.
Area having a block size of 16 × 16 pixels (horizontal 16 pixels × vertical 16 lines) for the luminance signal Y and two color difference signals C
An area consisting of an area having a block size of 8 × 8 pixels (horizontal 8 pixels × vertical 8 lines) for each of r and Cb (a macroblock quantization characteristic value (or macroblock) Quantization scale) QS},
It is a value indicated by the product of the quantization matrix.

Each of FIGS. 5 (b) to 5 (e) is a block of 8 × 8 pixels (horizontal 8 pixels × vertical 8 lines) which is a “unit block” in the MPEG1 system and the MPEG2 system. A pixel block having a size, and each (a) diagram in FIGS. 5 and 6 is composed of four “unit blocks” shown in each (b) to (e) diagram in FIG. 5 and FIG. 5 shows one area (macroblock), and one area described above is represented by numbers 1 to 16 in the horizontal direction and numbers 1 to 16 in the vertical direction in FIGS. 16 x 16 pixels (16 horizontal pixels x 16 vertical pixels
Pixel block having a block size of (line). The part delimited by the alternate long and short dash line in each of FIG. 5A is composed of 8 × 8 pixels (8 horizontal pixels × 8 vertical lines) shown in FIGS. 5B to 5E. It corresponds to a "unit block" having a block size. The correspondence between the blocks will be apparent from the numbers shown in the horizontal direction and the numbers shown in the vertical direction.

In FIGS. 5 and 6, subscripts o and e added to the numbers in the vertical direction indicate pixel columns (scanning) in odd fields when one image is formed by interlaced scanning of 2: 1. Line) and a pixel column (scanning line) in an even field are distinguished, and a pixel column (scanning line) in an odd field is attached with a subscript o, and a pixel column (scanning line) in an even field is attached. ) Subscript e is added, and in FIGS. 5 and 6, the drawing method in which a thin horizontal line is drawn for a pixel column (scanning line) in an odd field is adopted. In the MPEG1 system, only the image by the progressive scanning system is targeted for high efficiency encoding. Therefore, when FIG. do it. The meanings of horizontal and vertical numbers and subscripts attached to the vertical numbers in the first to third regions shown in FIGS. 3 and 4, respectively, are the same as those in FIGS. 5 and 6.
Is the same as that described above.

Now, the above-mentioned MPEG1 system, MPEG
In 2 methods, 8 x 8 pixels (8 horizontal pixels x 8 vertical lines)
Each of the 64 DCT transform coefficients for each block of each unit obtained by the DCT performed for each "unit block" of the block size is quantized by the "block quantization width value". The DCT coefficient which is quantized with the DCT transform coefficient as the dividend and the "block quantization width value" as the divisor is the DC component (DC component) of the DCT coefficient.
And an AC component (AC component). DC above
The DC component of the T coefficient is differentially encoded, and the AC component of the DCT coefficient is zigzag-scanned and then entropy-encoded (compressing the amount of information due to the deviation of the appearance probability of the code ... For example, variable-length encoding such as the Huffman method). The converted and encoded image data includes additional information necessary for decoding the transformed and encoded image data [for example, {block quantization width information → macroblock quantization characteristic value (or macroblock (Quantization scale) QS}, motion vector, prediction mode information, and an image of a frame structure in which a frame is an encoding unit and an image of a field structure in which a field is an encoding unit. (Picture structure) information, orthogonal transformation mode information, etc.] is added to form a coded bit stream. In the image data decoding apparatus of the present invention shown in FIGS. 1 and 2, the encoded bit stream supplied to the input terminal 1 is configured by using, for example, a first-in first-out memory (FIFO). It is stored in the buffer memory 8.

In the variable length decoding unit 9 to which the coded bit stream read from the buffer memory 8 is supplied, the entropy coded (variable length coded) image data and the transform coded image data are stored. Additional information required at the time of decoding, for example, {block quantization width information → macroblock quantization characteristic value (or macroblock quantization scale) QS}, motion vector, prediction mode information,
The image structure information (picture structure) information, the orthogonal transformation mode information, and the like are decoded. Then, the image data decoded by the variable length decoding unit 9, block quantization width information {macroblock quantization characteristic value (or macroblock quantization scale) QS}, image structure (picture structure) information, The orthogonal transform mode information is supplied to the inverse quantization unit 10, and the motion vector, the prediction mode information, and the like are supplied to the inverse motion compensation unit 13.

Image data decoded by the variable length decoding unit 9 and block quantization width information {macroblock quantization characteristic value (or macroblock quantization scale) QS} in the decoded additional information, image In the dequantization unit 10 to which the information of the structure (picture structure) and the mode information of the orthogonal transformation are given, the DC obtained by performing the dequantization operation.
The T transform coefficient is supplied to the inverse orthogonal transform unit (inverse DCT) 11. The inverse orthogonal transform unit (inverse DCT) 11 performs a two-dimensional inverse DCT for each unit block, inversely transforms image data in the frequency domain into image data in the time domain, and supplies it to the addition unit 12. To do. The image data in the time domain supplied to the adding unit 12 as described above is an image in a state in which motion compensation is performed by the motion compensation unit 13 according to a coding type indicating a difference between intraframe coding and interframe coding. Output data is stored in the image memory 14 as output image data, with or without addition to the data. And FIG.
In the image data decoding apparatus of the present invention shown in each drawing of FIG. 2, the image data from the image memory 14 in the decoder integrated circuit 3 is output to the output terminal 2 via the low pass filter (low pass filter) 7. Has been output to.

High-efficiency coding of moving picture information by the MPEG1 system and the MPEG2 system is a combination of intra-frame coding by two-dimensional discrete cosine transform (two-dimensional DCT) and inter-frame coding, and motion compensation prediction is performed. The image signal of each screen, which is subjected to high-efficiency encoding and is subjected to inter-frame prediction, has a block size of 8 × 8 pixels (horizontal 8 pixels × vertical 8 lines). It is as described above that the DCT is performed for each "unit block" of each unit, and the DCT is performed for each unit block.
In the PEG1 method, only moving image information obtained by applying the progressive scanning method is targeted for high-efficiency encoding.
In the EG2 method, moving image information obtained by applying the interlaced scanning method is also a target of high efficiency encoding.

Therefore, as the DCT performed in the high-efficiency encoding according to the MPEG2 system, as in the DCT performed in the high-efficiency encoding according to the MPEG1 system, the screen of each one of the screens subjected to the high-efficiency encoding is displayed. Orthogonal transform mode in which DCT is performed for each unit block obtained by dividing the image signal into “unit blocks” having a block size of 8 × 8 pixels (8 pixels horizontally × 8 lines vertically) 1 orthogonal transformation mode) is performed, and image data in each of the above-described areas is executed for each area (macroblock) of a predetermined size including at least one block of the above-described unit. The first orthogonal transformation mode in which the orthogonal transformation performed is performed on the image data for each block of the unit described above, and the interlacing in the image data in the region described above. The image data of a block of a new unit having the above-mentioned predetermined block size obtained by combining the image data of the odd field of scanning and the image data of the even field of the interlaced scanning in the image data in the area are combined. The image data of the block of the new unit having the predetermined block size obtained as described above is also subjected to any one of the second orthogonal transformation mode in which the orthogonal transformation is performed.

DC in the first orthogonal transform mode described above
Each of the "unit blocks" shown in FIGS. 5B to 5E exemplifying the "unit block" in which T is performed.
And an area (macroblock) shown in FIG. 5A configured to include each of the “unit blocks”, that is, an area set in the image targeted for high efficiency encoding. 5 is clear from the corresponding relationship between the horizontal numeral string and the vertical numeral string shown in each figure of FIG. In the MPEG1 system, moving image information obtained by applying the progressive scanning system is targeted for high efficiency encoding.
The subscripts o and e added to the numbers in the vertical number string shown in FIG. 5 are unnecessary.

Further, each new "unit block" shown in (b) to (e) of FIG. 6 exemplifying the "unit block" in which the DCT in the second orthogonal transform mode is performed. , An area (macroblock) shown in (a) of FIG. 6 that is configured to include "unit blocks", that is, an image of an odd field of interlaced scanning in the image data in the area shown in (a) of FIG. Image data of a new unit block {(b), (c) in FIG. 6) having a predetermined block size obtained by combining data, and an even field of interlaced scanning in the image data in the region described above. The relationship with the image data of a new block {(d), (e) in FIG. 6) having a predetermined block size obtained by combining the image data in FIG. Being sideways It evidenced by correspondence between the number string and the longitudinal direction of the string of digits.

Additional information required for decoding, which is added to the transform-coded image data to be decoded [for example, {block quantization width information → macroblock quantization characteristic value (or macro Block quantization scale) QS}, motion vector, prediction mode information, picture structure (picture structure) information, orthogonal transformation mode information, etc.] After being decoded by the unit 9, the above-mentioned respective additional information is supplied to each component in the decoder integrated circuit 3 and used for controlling the operation of each component. Among the additional information decoded by the modulatable decoding unit 9 provided in the decoder integrated circuit 3, information such as image structure information (picture structure) information and orthogonal transformation mode information is the decoder integrated circuit 3 In the case where the decoder integrated circuit 3 is configured so that it can be extracted from the above, as in the image data decoding apparatus shown in FIG.
The information of the picture structure (picture structure) and the mode information of the orthogonal transformation, which have been decoded in (3), are taken out from the decoder integrated circuit 3 and supplied to the control signal generating section 6.

In addition, in the additional information decoded by the modulatable decoding unit 9 provided in the decoder integrated circuit 3, information such as image structure (picture structure) information and orthogonal transform mode information is When the decoder integrated circuit 3 is configured in a state in which it cannot be taken out from the decoder integrated circuit 3,
Like the image data decoding apparatus shown in FIG. 1, the detection unit for detecting the image structure information and the orthogonal transformation mode information, which has the same function as the modulatable decoding unit 9 provided in the decoder integrated circuit 3 described above. 5, the encoded bit stream supplied to the input terminal 1 is, for example, a first-in first-out memory (FI
The image structure information and the orthogonal transformation mode information are supplied to the detection unit 5 via the buffer memory 4 configured using FO). Therefore, in the case of the image data decoding apparatus shown in FIG. 1, information on the image structure (picture structure) output from the detection unit 5 of the image structure information and the orthogonal transformation mode information and the orthogonal transformation mode information. And are supplied to the control signal generator 6.

As described above, the control signal generator 6 in the image data decoding apparatus shown in FIGS. 1 and 2 to which the image structure information and the orthogonal transformation mode information are supplied is the memory 15
The control unit 16, the Fi address generation unit 17, the Fr address generation unit 18, and the changeover switch 19 are provided. The image structure information and the orthogonal transformation mode information supplied to the control signal generator 6 are given to the controller 16 and stored in the memory 15. In addition, the control unit 1
An image address is also supplied to 6. When the image structure information supplied to the control unit 16 is the field structure Fi, the control unit 16 outputs a switching control signal for switching the movable contact v of the changeover switch 19 to the fixed contact i side. When the information of the image structure information that is generated and is supplied to the control unit 16 is the frame structure Fr, the control unit 16 changes the movable contact v of the changeover switch 19 to the fixed contact i.
A switching control signal is generated so as to switch to the side.

The image structure information supplied to the control unit 16 is the field structure Fi, and the movable contact v of the changeover switch 19 is switched to the fixed contact i side by the changeover control signal generated by the control unit 16. In the image memory 1, the control signal generator 6 outputs the Fi address signal generated by the Fi address generator 17 under the control of the controller 16.
4 and the decoded image data read from the image memory 14 is supplied to the low-pass filter 7, the low-pass filter 7 compares the decoded image data supplied thereto with
In the mode as illustrated in FIG. 3, the image of the field structure Fi is filtered in the vertical direction and output to the output terminal 2.

FIG. 3 illustrates a state of filtering processing in the vertical direction of an image of the field structure Fi when a digital filter having a filter length of 3 taps is used as the low pass filter 7. It is a figure used for. FIG. 3 shows three vertically continuous areas in a part of one image (first area in the figure,
The second area and the third area are shown. In FIG. 3, a set of pixel data designated by adding white circles → black circles → white circles in the vertical direction, for example, [[first area (horizontal 1, vertical 1o)]
Pixel data, and (horizontal 1, vertical 3o) pixel data,
(Set of (horizontal 1, vertical 5o) pixel data], [(horizontal 10, vertical 2e) pixel data in the first region, and (horizontal 1)
0, vertical 4e) pixel data and (horizontal 10, vertical 6e) pixel data set] ... [[horizontal 1 in the second region]
3, horizontal 16e) pixel data, (horizontal 13, vertical 2e) pixel data in the third area, (horizontal 13, vertical 4e)
e) a set of pixel data], [(horizontal 16, vertical 1o) pixel data, (horizontal 16, vertical 3o) pixel data, and (horizontal 13, vertical 5o) pixel data in the third region] The set] indicates a set of pixel data to be multiplied by three filter coefficients in a digital filter having a filter length of the FIR filter of 3 taps.

The set of three pixel data indicated by the above-mentioned white circles → black circles → white circles is [pixel data of (horizontal 1, vertical 1o) in the first area, and (horizontal 1) , Vertical 3
a combination of (o) pixel data and (horizontal 1, vertical 5 o) pixel data] → [(horizontal 2, vertical 1 o) pixel data and (horizontal 2, vertical 3 o) pixel data in the first region And a set of (horizontal 2, vertical 5o) pixel data] → [(horizontal 3, vertical 1o) pixel data in the first region, (horizontal 3, vertical 3o) pixel data, (horizontal 3 , Vertical 5o) pixel data]
→ [(horizontal 4, vertical 1o) pixel data in the first area, (horizontal 4, vertical 3o) pixel data, (horizontal 4, vertical 5o)
o) Pair with pixel data] →, the pixels move in the horizontal direction by one pixel each, and after reaching the position of the final pixel in the horizontal direction, shift by one pixel in the vertical direction. After moving from the first position in the horizontal direction in the state where the pixels are shifted by one pixel in the horizontal direction in the same manner as described above, and reaching the position of the final pixel in the horizontal direction, only one pixel in the vertical direction. It moves from the initial position in the lateral direction in the shifted state in the same state as described above. Then, in FIG. 3, a set of three pixel data with white circles → black circles → white circles is designated only for the positions of a small number of pixel data required for explanation. As for the mode of movement of the set of three pixel data as described above, and the point that only a part of the set of three pixel data marked with a white circle → a black circle → a white circle is designated, the same applies to FIG. It is the same.

As described above, the image structure information supplied to the control unit 16 is that of the field structure Fi.
In the state where the movable contact v of the changeover switch 19 is switched to the fixed contact i side by the changeover control signal generated in step 3, the pixel data read from the image memory 14 is displayed in the state as illustrated in FIG. Under the control of the control unit 16, the Fi address generating unit 1 controls the Fi address such that the decoded image data in such a state that the low-pass filter is applied in the vertical direction is sequentially supplied to the low-pass filter 7.
7 is supplied to the image memory 14 via the fixed contact i and the movable contact v of the changeover switch 19 so that the decoded image data read from the image memory 14 is supplied to the low-pass filter 7. Then, the low-pass filter 7
In FIG. 3, the decoded image data is output to the output terminal 2 in a state in which the image having the field structure Fi is filtered in the vertical direction in a manner as illustrated in FIG.

As described above, the movable contact v of the changeover switch 19 is brought into the state of being switched to the fixed contact i side.
Of the decoded additional information, in addition to the above-mentioned case where the picture structure (picture structure) shows the field structure Fi and the case where the orthogonal transformation mode information is the field DCT, the decoded additional information is However, the switching control signal is given from the control unit 16 to the changeover switch 19 so that the movable contact v of the changeover switch 19 is forcibly switched to the fixed contact i side. There is a case where the operation of the control unit 19 is controlled from a non-operating unit. As described above, regardless of the decoded additional information, the low-pass filter processing similar to the case where the image structure (picture structure) indicates the field structure Fi is always performed on the decoded image data. If such a measure is taken, for example, when the filtering process corresponding to the frame structure is performed on the part of the image content having a large motion, the problem that the resolution in the time direction is deteriorated is not caused. The advantage of being able to achieve a reduction in quantization noise is obtained.

Next, the image structure information supplied to the control unit 16 is the frame structure Fr, and the movable contact v of the changeover switch 19 is switched to the fixed contact r side by the changeover control signal generated by the control unit 16. In this state, the control signal generation unit 6 supplies the Fr address signal generated by the Fr address generation unit 18 to the image memory 14 under the control of the control unit 16, and the decoding read from the image memory 14 is performed. When the image data is supplied to the low-pass filter 7, the low-pass filter 7 compares the decoded image data supplied to the low-pass filter 7 with, for example, 1o in the vertical direction in the first region in FIG.
In the mode as illustrated in the range of 15o, the image of the frame structure Fr is filtered in the vertical direction and output to the output terminal 2.

When a digital filter having a filter length of 3 taps is used as the low-pass filter 7, the filter processing state in the vertical direction of the image of the field structure Fi is, for example, as shown in FIG. 1
This is as illustrated in the range from 1o to 15o in the vertical direction in the area. Although FIG. 4 shows three vertically continuous regions (a first region, a second region, and a third region in the figure) in a part of one image, It is assumed that both the first region and the third region are regions of the first orthogonal transform mode, and the second region is a region of the second orthogonal transform mode.

In the set of pixel data indicated by adding white circles → black circles → white circles in the vertical direction in FIG. 4, for example, [first
Group of (horizontal 2, vertical 1o) pixel data, (horizontal 2, vertical 3o) pixel data, and (horizontal 2, vertical 3o) pixel data] in the area 5, vertical 2
e) Pixel data, (horizontal 5, vertical 3o) pixel data, and (horizontal 50, vertical 4e) pixel data]] ...
[(Horizontal 2, vertical 12e) pixel data in the first region, (horizontal 2, vertical 13o) pixel data, (horizontal 2, vertical 1o)
4e) pixel data], [(horizontal 4, vertical 13o) pixel data in the first region, and (horizontal 4, vertical 14)
e) a set of pixel data of (horizontal 4, vertical 15 o)] is a set of pixel data multiplied by three filter coefficients in a digital filter having a filter length of 3 taps. With respect to the above range, the low-pass filter processing in the vertical direction of the image is performed in a form suitable for the image data of the frame structure Fr.

As described above, in FIG. 4, the first region and the third region
It is said that both the area and the area of 1 are the areas of the first orthogonal transformation mode, and the second area is the area of the second orthogonal transformation mode. When the successive regions (macroblocks) are adjacent to the regions having different orthogonal transform modes, the low-pass filter processing is performed in the manner as illustrated in FIG. 4, respectively. . First, looking at the vicinity of the boundary between the first region and the second region in FIG. 4, the first orthogonal region in the first region, which is the first orthogonal transform mode, is continuous in the vertical direction up to 15o. The pixel data in the state of arranging three of them is multiplied by the filter coefficient. However, the last one of the pixel data in the state where three pixels are arranged in succession in the vertical direction is located at the boundary between the first region and the second region, and 14e in the vertical direction in the first region.
For pixel data in a state where three pixels are continuously arranged in the vertical direction from the above, [pixel data of (horizontal 6, vertical 14 e) in the first region and (horizontal 6, vertical 16 e) in the first region are illustrated.
e) the pixel data and (horizontal 6, vertical 2) in the second area.
e) set with pixel data].

Pixel data near the boundary between the first region, which is said to be the first orthogonal transform mode region, and the second region, which is said to be the second orthogonal transform mode, is originally Also in the first region which is the region of the first orthogonal transformation mode where the filter coefficient should be multiplied to the pixel data in the state where three pixels are arranged consecutively in the vertical direction,
As for the pixel data in the portion close to the second region, the filter data is multiplied by the pixel data in a state where every other three pixel data are arranged in the vertical direction. Further, regarding the vicinity of the boundary between the second region and the third region in FIG. 4, every other pixel in the second region, which is the second orthogonal transform mode, starts in the vertical direction. A set of three pieces of pixel data in a state of arranging three pieces is the third orthogonal transformation mode, that is, the third set of pixel data.
The area is set to a range up to 4e in the vertical direction.

In the control signal generation unit 6, the orthogonal transformation mode information of the sequential regions arranged in the vertical direction of the image read from the memory 15 which stores the orthogonal transformation mode information of the plurality of regions arranged in the vertical direction of the image. And the Fi address and Fr address generated based on the pixel address, the information of the image structure, etc. are selectively supplied to the image memory 14 by the changeover switch 19, and the optimum vertical for each area set in the image. Enables low-pass filtering of directions.

The low-pass filter processing to be performed on the decoded image data in the vertical direction of the image is performed using every other vertical decoded pixel data in the block of each unit. In the case of the cutoff frequency of the low-pass filter, the low-pass filter processing to be performed on the decoded image data in the vertical direction of the image is performed using the decoded pixel data that is continuous in the vertical direction in each unit block. It is better to set it higher than the cut-off frequency of the low-pass filter in the case that the low-pass filter is to be applied to the decoded image data in the vertical direction of the image. Cutoff characteristic curve of low-pass filter when decoding is performed using every other decoded pixel data In the case where the line inclination is lower than the low-pass filter processing to be performed on the decoded image data in the vertical direction of the image using the decoded pixel data that is continuous in the vertical direction in each unit block, It is preferable that the cutoff characteristic curve of the low-pass filter is set to be gentler than the slope.

[0046]

As is apparent from the above description in detail, according to the present invention, one of an image having a frame structure in which a frame is an encoding unit and an image in a field structure having a field as an encoding unit are used. Alternatively, a predetermined block size (N × M pixels ← horizontal N pixels ×) is set for the image signal of each screen that is the target of high efficiency encoding by both.
For each “unit block” having a block size of vertical M lines), (N × M) orthogonal transform coefficients obtained by orthogonal transform are predetermined including at least one of the unit blocks described above. Transform-encoded image data obtained by performing predetermined encoding after being quantized by the “block quantization width value” set for each size region, and the transform-encoded image Image structure information (picture structure) information detected from the additional information included in the bitstream by the additional information required when decoding the data (indicating whether the image structure is a frame structure or a field structure) Information), orthogonal transform mode information, pixel address, etc., and a low-pass filter process to be performed on the image data in the vertical direction of the image is set in the device. The address signal necessary for causing the output terminal 2 to output the decoded image data in the state performed by the low-pass filter (low-pass filter) according to the selected operation mode is supplied to the image memory, In one of the operation modes to be set, low-pass filter processing to be performed in the vertical direction of the image and decoded second orthogonal For the decoded image data in the area that had been orthogonally transformed by the transformation mode, both the low-pass filtering process and the low-pass filtering process that should be performed in the vertical direction of the image are the low-pass filtering process for the image data of the field of the image by the interlaced scanning method. Or the orthogonal transformation is performed by the first orthogonal transformation mode. Regarding the decoded image data corresponding to the unit block that has been made, in the state where the decoded image data selected for the low-pass filter processing in the vertical direction of the image does not exist across the unit block boundary. , The low-pass filter processing to be performed on the decoded image data in the vertical direction of the image is performed using the decoded pixel data that is continuous in the vertical direction in the unit block, or Regarding the decoded image data corresponding to the block of the unit that has been orthogonally transformed by the orthogonal transformation mode, the decoded image data selected for the low-pass filter processing in the vertical direction of the image has a unit block boundary. In the state that does not exist in
The low-pass filter processing to be performed on the decoded image data in the vertical direction of the image is performed by using every other vertical decoded pixel data in the unit block, and further, in the vertical direction of the image. In a state in which the decoded image data selected for the low-pass filter processing for the above exists in two unit blocks that are adjacent to each other across the boundary of the unit block, Both of the unit blocks are first
In the case of the block of the unit of the decoded image data corresponding to the block of the unit that has been orthogonally transformed by the orthogonal transformation mode of, the decoded image data should be applied in the vertical direction of the image. The low-pass filter processing may be performed using decoded pixel data that is continuous in the vertical direction in each unit block,
Furthermore, in a state where the decoded image data selected for the low-pass filter processing in the vertical direction of the image exists in two unit blocks that are adjacent to each other across the unit block boundary, the above-mentioned boundary is used. In the case where at least one of the two unit blocks adjacent to each other across the block is a unit block of decoded image data corresponding to the unit block that has been orthogonally transformed in the second orthogonal transform mode, The low-pass filter processing to be performed on the decoded image data of in the vertical direction of the image is 1 in the vertical direction in each unit block.
By performing low-pass filter processing on image data using alternate decoded pixel data, block noise and quantization noise such as mosquito noise that occur during image decoding by the interlaced scanning method can be satisfactorily reduced. According to the present invention, the above-mentioned problems can be easily solved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration example of a decoding device for transform-coded image data by applying a method of reducing quantization noise generated when decoding transform-coded image data according to the present invention.

FIG. 2 is a block diagram showing an example of the configuration of a decoding apparatus for image data that has been transform-coded by applying the method for reducing quantization noise that occurs when decoding transform-coded image data according to the present invention.

FIG. 3 is a diagram for explaining the operation of the low-pass filter.

FIG. 4 is a diagram for explaining the operation of the low-pass filter.

FIG. 5 is a diagram for explaining a block of pixels.

FIG. 6 is a diagram for explaining a block of pixels.

[Explanation of symbols]

1 ... Input terminal of encoded bit stream (encoded data string) to be decoded, 2 ... Output terminal, 3
... component parts integrated as a decoder integrated circuit, 6 ... control signal generation part, 7 ... low-pass filter, 8 ... buffer memory, 9 ... variable length decoding part, 10 ... dequantization part, 11 ... Inverse orthogonal transformation unit, 12 ... Addition unit, 13 ... Inverse motion compensation unit, 14 ... Image memory, 15 ... Memory, 16 ... Control unit, 17 ... Fi address generation unit, 18 ... Fr address generation unit, 19 ... Switching switch,

Claims (8)

[Claims] 1. Image data to be transform-encoded by one or both of an image of a frame structure having a frame as an encoding unit and an image of a field structure having a field as an encoding unit, The image data in each of the regions is divided into blocks of a unit having a predetermined block size, and the region of a predetermined size including at least one of the blocks of the units described above. The orthogonal transformation to be performed is, for an image having a frame structure, the first orthogonal transformation mode which is performed for the image data of each block of the above-mentioned unit and the image data of the odd field of the interlaced scanning in the image data in the above-mentioned area Image data of a new unit block having a predetermined block size obtained by combining With the second orthogonal transform mode in which the orthogonal transform is performed on the image data of the new unit block having the predetermined block size obtained by combining the image data of the even field of the interlaced scanning in the image data of In contrast to any of the orthogonal transform modes, an image of a field structure is adapted to be performed in the second orthogonal transform mode described above. A method of reducing quantization noise that occurs when decoding image data that has been quantized by transform coding by a predetermined coding while being quantized using a block quantization width value that is individually set to The decoded image data in the area that has been orthogonally transformed by the first orthogonal transformation mode described above is applied in the vertical direction of the image. Power low-pass filter processing and low-pass filter processing to be performed in the vertical direction of the image on the decoded image data of the area that has been orthogonally transformed in the second orthogonal transformation mode described above, as an interlaced scan. A method for reducing quantization noise that occurs when decoding transform-encoded image data by applying a low-pass filter process to image data in a field of an image by a method. 2. Image data to be transform-encoded by one or both of an image having a frame structure having a frame as an encoding unit and an image having a field structure having a field as an encoding unit, The image data in each of the regions is divided into blocks of a unit having a predetermined block size, and the region of a predetermined size including at least one of the blocks of the units described above. The orthogonal transformation to be performed is, for an image having a frame structure, the first orthogonal transformation mode which is performed for the image data of each block of the above-mentioned unit and the image data of the odd field of the interlaced scanning in the image data in the above-mentioned area. Image data of a new unit block having a predetermined block size obtained by combining With the second orthogonal transform mode in which the orthogonal transform is performed on the image data of the new unit block having the predetermined block size obtained by combining the image data of the even field of the interlaced scanning in the image data of In contrast to any of the orthogonal transform modes, an image of a field structure is adapted to be performed in the second orthogonal transform mode described above. In order to decode the image data that has been quantized by using the block quantization width value that is individually set to, and at least the buffer memory and the variable length decoding unit And a decoding device for transform-coded image data, which includes an inverse quantization unit, an inverse orthogonal transformation unit, and an image memory. I, from the transform encoded image data, the size of each area predetermined comprising one block of units by at least the,
The orthogonal transformation performed on the image data in each of the regions described above is performed in the first orthogonal transformation mode performed on the image data of each block of the unit described above, and the odd field of the interlaced scanning in the image data in the regions described above. Image data of a block of a new unit having a predetermined block size obtained by combining the image data of and image data of an even field of interlaced scanning in the image data in the area Orthogonal transform mode information and image structure indicating whether any one of the second orthogonal transform mode in which the orthogonal transform is performed and the image data of the new unit block having the predetermined block size is performed. Image structure information and orthogonal transformation mode information detecting means for detecting information, and reading from the image memory By the low-pass filter that passes the decoded image data and the detection power of the image structure information and the orthogonal transformation mode information detecting means, the low-pass filter always performs the low-pass filtering process on the image data of the field of the image by the interlaced scanning method. And a decoding device for transform-encoded image data. 3. Image data to be transform-encoded by one or both of an image having a frame structure having a frame as an encoding unit and an image having a field structure having a field as an encoding unit, The image data in each of the areas is divided into blocks of a unit having a predetermined block size, and for each area of a predetermined size including at least one of the blocks of the units described above. The orthogonal transformation to be performed is, for an image having a frame structure, the first orthogonal transformation mode which is performed for the image data of each block of the unit described above and the image data of the odd field of the interlaced scanning in the image data in the area. Image data of a new unit block having a predetermined block size obtained by combining With the second orthogonal transform mode in which the orthogonal transform is performed on the image data of the new unit block having the predetermined block size obtained by combining the image data of the even field of the interlaced scanning in the image data of In contrast to any of the orthogonal transform modes, an image of a field structure is adapted to be performed in the second orthogonal transform mode described above. In order to decode the image data that has been quantized by using the block quantization width value that is individually set to, and at least the buffer memory and the variable length decoding unit And a decoding device for transform-coded image data, which includes an inverse quantization unit, an inverse orthogonal transformation unit, and an image memory. Then, based on the image structure information and the orthogonal transformation mode information detected by the variable length decoding unit, the decoded image read from the image memory based on the low-pass filter that passes the decoded image data read from the image memory. A device for decoding conversion-coded image data, comprising a means for allowing a low-pass filter for passing data to always perform low-pass filter processing on image data of an image field by an interlaced scanning method. 4. Image data to be transform-encoded by one or both of an image of a frame structure having a frame as an encoding unit and an image of a field structure having a field as an encoding unit, The image data in each of the regions is divided into blocks of a unit having a predetermined block size, and the region of a predetermined size including at least one of the blocks of the units described above. The orthogonal transformation to be performed is, for an image having a frame structure, the first orthogonal transformation mode which is performed for the image data of each block of the above-mentioned unit and the image data of the odd field of the interlaced scanning in the image data in the above-mentioned area. Image data of a new unit block having a predetermined block size obtained by combining A second orthogonal transformation mode in which orthogonal transformation is performed on the image data of a new unit block having a predetermined block size obtained by combining the image data of the even field of the interlaced scanning In contrast to any of the orthogonal transform modes, an image of a field structure is adapted to be performed in the second orthogonal transform mode described above. A method of reducing quantization noise that occurs when decoding image data that has been quantized by transform coding by a predetermined coding while being quantized using a block quantization width value that is individually set to Regarding the low-pass filter processing performed on the decoded image data in the vertical direction of the image, the first orthogonal transform mode is used. Regarding the decoded image data corresponding to the unit block for which the orthogonal transformation has been performed, the decoded image data selected for the low-pass filter processing in the vertical direction of the image exists across the boundary of the unit block. In the non-existence state, the low-pass filter processing to be performed on the decoded image data in the vertical direction of the image is performed using the decoded pixel data that is continuous in the vertical direction in the unit block, and Regarding the decoded image data corresponding to the block of the unit that has been orthogonally transformed by the orthogonal transformation mode of No. 2, the decoded image data selected for the low-pass filter processing in the vertical direction of the image has the boundary of the unit block. The vertical direction of the image with respect to the above-mentioned decoded image data in the state where it does not exist in between. The low-pass filter processing to be performed on the horizontal direction is performed using every other decoded pixel data in the vertical direction in the unit block, and the decoded image data selected for the low-pass filter processing on the vertical direction of the image. However, in the state in which the blocks of two units adjacent to each other across the boundary of the unit block are present, both blocks of the two units adjacent to each other across the boundary are orthogonal by the first orthogonal transformation mode. In the case of the block of the unit of the decoded image data corresponding to the block of the unit that has been converted, the low-pass filtering process to be performed on the decoded image data in the vertical direction of the image is Is performed by using decoded pixel data that is continuous in the vertical direction in a unit block, and also
In a state where the decoded image data selected for the low-pass filter processing in the vertical direction of the image is present in two unit blocks that are adjacent to each other across the boundary of the unit block, they are adjacent to each other across the boundary. In the case where at least one of the two matching unit blocks is the unit block of the decoded image data corresponding to the unit block that has been orthogonally transformed in the second orthogonal transform mode, it is decoded as described above. The low-pass filter processing that should be applied to the image data in the vertical direction of the image,
A method for reducing quantization noise generated when decoding transform-coded image data, which is performed using every other vertical decoded pixel data in a block of each unit. 5. Image data to be transform-encoded by one or both of a frame-structured image in which a frame is an encoding unit and a field-structured image in which a field is an encoding unit, The image data in each of the regions is divided into blocks of a unit having a predetermined block size, and the region of a predetermined size including at least one of the blocks of the units described above. The orthogonal transformation to be performed is, for an image having a frame structure, the first orthogonal transformation mode which is performed for the image data of each block of the above-mentioned unit and the image data of the odd field of the interlaced scanning in the image data in the above-mentioned area. Image data of a new unit block having a predetermined block size obtained by combining With the second orthogonal transform mode in which the orthogonal transform is performed on the image data of the new unit block having the predetermined block size obtained by combining the image data of the even field of the interlaced scanning in the image data of In contrast to any of the orthogonal transform modes, an image of a field structure is adapted to be performed in the second orthogonal transform mode described above. In order to decode the image data that has been quantized by using the block quantization width value that is individually set to, and at least the buffer memory and the variable length decoding unit And a decoding device for transform-coded image data, which includes an inverse quantization unit, an inverse orthogonal transformation unit, and an image memory. I, from the transform encoded image data, the size of each area predetermined comprising one block of units by at least the,
The orthogonal transformation performed on the image data in each of the regions described above is performed in the first orthogonal transformation mode performed on the image data of each block of the unit described above, and the odd field of the interlaced scanning in the image data in the regions described above. Image data of a block of a new unit having a predetermined block size obtained by combining the image data of and image data of an even field of interlaced scanning in the image data in the area Orthogonal transform mode information and image structure indicating whether any one of the second orthogonal transform mode in which the orthogonal transform is performed and the image data of the new unit block having the predetermined block size is performed. Image structure information and orthogonal transformation mode information detecting means for detecting information, and reading from the image memory Corresponds to a low-pass filter that passes the decoded image data that has been passed through, and a unit block that has undergone orthogonal transformation in the first orthogonal transformation mode based on the detection power of the image structure information and orthogonal transformation mode information detection means described above. As for the decoded image data to be decoded, the decoded image data selected in the vertical direction of the image in the state where the decoded image data selected for the low-pass filtering process does not exist across the boundary of the unit block. , The low-pass filter processing to be performed in the vertical direction of the image is performed using the decoded pixel data that is continuous in the vertical direction in the unit block, and the orthogonal transformation is performed in the second orthogonal transformation mode. For the decoded image data corresponding to the block of the In the state in which the decoded image data selected for the low-pass filtering does not exist across the block boundaries of the unit,
The low-pass filtering process to be performed on the decoded image data in the vertical direction of the image is further performed so that every other decoded pixel data in the vertical direction in the unit block is used. Decoded image data selected for the low-pass filter process in the vertical direction are adjacent to each other with a unit block boundary therebetween.
In the state of being present in the block of one unit, both of the blocks of two units which are adjacent to each other across the boundary correspond to the block of the unit in which the orthogonal transformation is performed in the first orthogonal transformation mode. In the case of a unit block of decoded image data, low-pass filter processing to be performed on the decoded image data in the vertical direction of the image is decoded pixel data that is continuous in the vertical direction in each unit block. Furthermore, the decoded image data selected for low-pass filtering in the vertical direction of the image exists in two unit blocks adjacent to each other across the boundary of the unit block. In this state, at least one of the blocks of two units adjacent to each other across the boundary is the second orthogonal transform module. In the case of the block of the unit of the decoded image data corresponding to the block of the unit which has been orthogonally transformed by the code, the low-pass filtering process to be performed on the decoded image data in the vertical direction of the image. However, a decoding device for transform-coded image data, comprising means for operating a low-pass filter so as to be performed by using every other vertical decoded pixel data in each unit block. 6. Image data to be transform-encoded by one or both of an image of a frame structure having a frame as an encoding unit and an image of a field structure having a field as an encoding unit, The image data in each of the areas is divided into blocks of a unit having a predetermined block size, and for each area of a predetermined size including at least one of the blocks of the units described above. The orthogonal transformation to be performed is, for an image having a frame structure, the first orthogonal transformation mode which is performed for the image data of each block of the above-mentioned unit and the image data of the odd field of the interlaced scanning in the image data in the above-mentioned area. Image data of a new unit block having a predetermined block size obtained by combining With the second orthogonal transform mode in which the orthogonal transform is performed on the image data of the new unit block having the predetermined block size obtained by combining the image data of the even field of the interlaced scanning in the image data of In contrast to any of the orthogonal transform modes, an image of a field structure is adapted to be performed in the second orthogonal transform mode described above. In order to decode the image data that has been quantized by using the block quantization width value that is individually set to, and at least the buffer memory and the variable length decoding unit And a decoding device for transform-coded image data, which includes an inverse quantization unit, an inverse orthogonal transformation unit, and an image memory. Then, based on the low-pass filter that passes the decoded image data read from the image memory and the image structure information and the orthogonal transformation mode information detected by the variable length decoding unit, the orthogonal transformation is performed by the first orthogonal transformation mode. Regarding the decoded image data corresponding to the unit block that has been converted, the decoded image data selected for the low-pass filter processing in the vertical direction of the image does not exist across the unit block boundary. In the state, the low-pass filter processing to be performed on the decoded image data in the vertical direction of the image is performed using the decoded pixel data that is continuous in the vertical direction in the unit block, and Decoded image data corresponding to a unit block that has been orthogonally transformed in the second orthogonal transformation mode As for the decoded image data selected for the low-pass filter processing in the vertical direction of the image, in the state where the decoded image data does not exist across the unit block boundary, The low-pass filtering to be performed in the vertical direction was selected for the low-pass filtering in the vertical direction of the image so that it is performed using every other vertical decoded pixel data in the unit block. In a state where the decoded image data is present in two unit blocks that are adjacent to each other across the boundary of the unit block, both of the two unit blocks that are adjacent to each other across the boundary are the first orthogonal. In the case of a unit block based on decoded image data corresponding to a unit block that has been orthogonally transformed according to the transform mode Further, so that the low-pass filter processing to be performed on the decoded image data in the vertical direction of the image is performed using the decoded pixel data that is continuous in the vertical direction in each unit block. Further, when the decoded image data selected for the low-pass filter processing in the vertical direction of the image is present in two unit blocks adjacent to each other across the unit block boundary, In the case where at least one of the two unit blocks adjacent to each other in between is the unit block of the decoded image data corresponding to the unit block for which the orthogonal transformation has been performed in the second orthogonal transformation mode, The low-pass filter processing that should be performed on the decoded image data in the vertical direction of the image is applied to each unit block. Decoder transform coding image data comprising a means for operating a low-pass filter as is done using the decoded pixel data of every other vertical that. 7. A case where low-pass filter processing to be applied to decoded image data in the vertical direction of an image is performed using every other vertical decoded pixel data in a block of each unit. The cut-off frequency of the low-pass filter is lower than the low-pass filter that should be applied to the decoded image data in the vertical direction of the image by using decoded pixel data that is continuous in the vertical direction in each unit block. The decoding device for transform-coded image data according to claim 5 or 6, wherein the cutoff frequency is set higher than the cutoff frequency of the low-pass filter. 8. A low-pass filter process to be applied to decoded image data in the vertical direction of an image is performed by using every other vertical decoded pixel data in a block of each unit. The slope of the cutoff characteristic curve of the low-pass filter should be applied to the decoded image data in the vertical direction of the image. The low-pass filter processing uses the decoded pixel data that is continuous in the vertical direction in each unit block. 8. The decoding apparatus for transform-coded image data according to claim 5, wherein the slope is set to be gentler than the slope of the cutoff characteristic curve of the low-pass filter.