JPH11331844A - Fixed length block encoder and decoder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve a picture quality without deteriorating encoding efficiency by controlling the code length of a coefficient value equivalent to one encoded block to be a fixed length. SOLUTION: A picture transformation part 302 applies two-dimensional discrete cosine transformation(DCT) to picture data xij by each rectangular block to transform to a 8×8 DCT coefficient value yuv. Then an encoding order deciding part 303 decides an order when a following successively encoding part 304 encodes the DCT coefficient value for each coefficient value yuv. The coefficient encoded by the part 304 based on this order grasps the inputted code length of each coefficient by a code length controlling part 305 to control the code length of each coefficient so that the whole of one block may be a code length fixed in advance. This part 305 adds coding order information, for example, to the head of the code string of one block, and transmits it to a decoding part. Consequently, a partial feature in a picture can be reflected on a code and the picture quality is improved without deteriorating encoding efficiency.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a fixed-length block encoding device and a decoding device for compressing multilevel image data.

[0002]

2. Description of the Related Art As a method of encoding a multi-valued image using a two-dimensional discrete cosine transform (DCT), ITU-
The JPEG system standardized by T and ISO is known. As another conventional example of an encoding method using DCT, there is proposed a method of processing a coefficient value by reducing one block as shown in, for example, Japanese Patent Application Laid-Open No. 2-122767.

[0003] As a coding / decoding system used in a conventional facsimile apparatus, when processing a binary image, a MH, MR, MMR system or a JBI using an arithmetic code is used.
There are known a G method and a JPEG method using adaptive discrete cosine transform (ADCT) when processing a multi-valued image. FIG. 8 shows a conventional encoding unit and decoding unit of the JPEG system using DCT. In the encoding unit, first, an image is read out by the block reading unit 201 for each block = 8 × 8 pixels, and then the image data xij of 8 × 8 pixels is converted to 8 × 8 pixels for each block by the DCT conversion unit 202 based on the following equation. Is converted to a DCT coefficient value yuv.

[0004]

(Equation 1)

Next, this DCT coefficient value yuv is
03, the coefficient is linearly quantized at a quantization step size having a different size for each coefficient yuv and is converted into a quantized coefficient. At this time, the quantization step size for each coefficient yuv is obtained by multiplying the quantization matrix stored in advance in the quantization matrix storage unit 206 by the scale factor set by the scale factor control unit 205. Also, by changing the value of the scale factor, it is possible to control the encoding amount and the quality of the decoded image. Next, the Huffman encoding unit 204 assigns a short code to a coefficient with a high appearance frequency and assigns a long code to a coefficient with a low appearance frequency, thereby compressing the code amount as a whole.

On the other hand, in the decoding unit, first, the coefficient quantized by the quantization unit 203 on the encoding unit side is reproduced by the Huffman decoding unit 207 based on the code, and then the quantized coefficient is inversely quantized. The DCT coefficient value yuv is inversely quantized by the transform unit 208. The quantization step size in the inverse quantization is the same as the encoding unit side, with respect to the quantization matrix stored in the quantization matrix storage unit.
It is obtained by multiplying the scale factor set by the scale factor control unit 205. This DC
The T coefficient value yuv is 1 block = 8 × by the inverse DCT transform unit.
The image data xij is inversely transformed into 8-pixel image data xij, and then restored by the block writing unit 210 in the original order.

In this example, the quantization matrix stored in the quantization matrix storage unit 206 is used in advance, but the one used in the encoding unit is transmitted to the decoding unit to store it in advance. Therefore, the quantization matrix storage unit 206 can be omitted as a result.

[0008]

In recent years, digital image devices such as copying machines, facsimile machines, filing devices and the like tend to increase the resolution or increase the number of gradations in order to improve image quality. . However, although the image quality is improved, the information amount of the image is increased. For example, when the image of two gradations (white and black) is increased to 256 gradations, the information amount becomes eight times. If the information amount is stored in the memory as it is, an eight-fold memory capacity is required and the cost increases. To prevent this, multi-valued image data must be compressed.

[0009] Image data encoding methods can be broadly classified into a variable length method and a fixed length method when classified according to the codeword length after encoding. The former has the features of higher encoding efficiency and reversibility compared to the latter,
Compared to the former, the latter is characterized in that the position of the image before encoding can be specified in the state of the code, so that only an arbitrary part of the image can be reproduced. Therefore,
From these characteristics, the former is suitable for a device such as a facsimile, and the latter is suitable for a device such as a copying machine that performs image processing.

Here, the simplest method as a fixed-length encoding method suitable for a copying machine or the like is a method of fixing a code word length assigned to one pixel. For example, for an image of 8 bits per pixel, This is a method of allocating 2 bits per pixel. However, this method has a problem that a partial characteristic of an image cannot be reflected on a code. In other words, it is not possible to distinguish an image region that is visually important from an image region that is not so important, so that the image is inferior in image quality. Further, in order to improve the image quality, the number of bits allocated to one pixel must be increased, and there is a problem that the coding efficiency is deteriorated.

The present invention has been made in view of the above-mentioned problems of the prior art, and when encoding an image using a fixed-length method, a fixed-length block encoding apparatus and a decoding method capable of improving image quality without deteriorating encoding efficiency. It is an object to provide a chemical conversion device.

[0012]

In order to achieve the above object, a first means is an image dividing means for dividing multi-valued image data into predetermined rectangular blocks, and a block divided by the image dividing means. Image conversion means for converting the image data in the image data into coefficient values by two-dimensional discrete cosine conversion, and encoding the important coefficients based on the coefficient values converted by the image conversion means so as to perform encoding first from important coefficients. Coding order determining means for determining an order; coding means for coding the coefficient values converted by the image converting means based on the coding order determined by the coding order determining means; Code length control means for controlling the code length of the coefficient value for one block encoded by the means to be a predetermined length.

The second means is the first means, wherein the coding order determining means selects one of a plurality of predetermined coding orders based on the coefficient value converted by the image converting means. It is characterized by the following.

A third means is that, when the code length of the coefficient value encoded by the encoding means exceeds the predetermined length, the code length control means in the first and second means, Is deleted.

A fourth means is that, in the first and second means, when the code length control means has a code length of the coefficient value coded by the coding means exceeding the predetermined length, the code length exceeds the predetermined length. Is converted to fit within the predetermined length and stored in the block.

A fifth means is that, in the first and second means, when the code length control means has a code length of a coefficient value coded by the coding means exceeding the predetermined length, the code length exceeds the predetermined length. Is converted to fit within the predetermined length, and when the difference between the code before and after conversion is equal to or more than a predetermined value, discard the code exceeding the predetermined value, and when the difference is equal to or less than the predetermined value, store the converted code in the block. It is characterized by.

The sixth means is characterized in that in the first and second means, the code length control means stores the information of the coding order determined by the coding order determining means in the block. .

The seventh means is a fixed-length block decoding device for decoding a code coded by the fixed-length block coding device of the sixth means, wherein the coefficient value coded by the coding means is provided. Decoding means for decoding based on the coding order determined by the coding order determining means,
Image inverse transform means for inversely transforming the coefficient value decoded by the decoding means into the image data for each block; and the image data for each block inversely transformed by the image inverse transform means to original order image data. Image reproducing means for reproducing.

[0019]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a facsimile apparatus to which an embodiment of an encoding apparatus and a decoding apparatus according to the present invention is applied, FIG. 2 is a block diagram showing an encoding unit of FIG. 1 in detail, and FIG. FIG. 4 is an explanatory diagram showing another example of the encoding order determined by the encoding order determination unit.
FIG. 5 is an explanatory diagram showing the code length control of the code length control unit of FIG.
FIG. 6 is an explanatory diagram showing another example of the code length control of the code length control unit, FIG. 6 is an explanatory diagram showing the format processing of the code length control unit in FIG. 2, and FIG. 7 is a block diagram showing the decoding unit in FIG. 1 in detail. is there.

Referring to FIG. 1, in the facsimile apparatus on the transmitting side, an original is read by a CCD image sensor or the like of an image reading unit 101, and then the read data is subjected to image processing by an image processing unit 102 so as to be suitable as transmission data. Then, the image data subjected to image processing by the image processing unit 102 is encoded by the encoding unit 103,
This encoded data is transmitted to the receiving facsimile machine via the transmission line 100. In the receiving-side facsimile apparatus, the decoding unit 104 decodes the received data into image data, and then this image data is sent to the image output unit 1 such as a plotter.
Image processing is performed by the image processing unit 105 so as to be suitable for the image processing unit 06, and then a hard copy is output by the image output unit 106. Examples of processing by the image processing units 102 and 106 include resolution conversion and size conversion for binary images, and color (color component) conversion, resolution conversion, and size conversion for multi-valued images containing colors. And so on.

Next, the encoding unit 103 according to the present invention will be described with reference to FIG. First, similarly to the conventional example, the block dividing unit 301 divides the image data into rectangular blocks of, for example, 8 × 8 pixels per block, and then the image converting unit 302 converts the image data xij into a two-dimensional DCT for each rectangular block. To obtain the 8 × 8 DCT coefficient value yuv
Convert to In the present invention, the encoding order determination unit 3
03, based on the distribution state of the DCT coefficient values yuv, the order in which the subsequent sequential encoding unit 304 encodes the DCT coefficient values yuv is determined for each coefficient value yuv.

As a method of determining the encoding order of the DCT coefficient value yuv, encoding of more important coefficients is performed first based on the magnitude of the energy of the AC coefficient in the coefficient value yuv, the deviation of the AC coefficient value, and the like. Do. Coefficients sequentially encoded by the encoding unit 304 based on this order are applied to the code length control unit 305. The code length control unit 305 grasps the input code length of each coefficient and determines one block as a whole. The code length of each coefficient is controlled so that the code length becomes the same.

Therefore, as described above, the DCT coefficient value yuv
The order in which the DCT coefficient values yuv are encoded is determined for each coefficient value yuv based on the distribution state, and then the encoding is performed. Then, the code length of each coefficient is set so that the entire code block has a predetermined code length. Since the control is performed, partial characteristics of the image can be reflected on the code, and as a result, the image quality can be improved without lowering the coding efficiency.

By the way, according to such processing, DC
Since the encoding order is determined for each coefficient value yuv based on the distribution state of the T coefficient value yuv, the highest image quality can be obtained at a predetermined code length, but the processing becomes complicated and the processing speed is reduced. Become slow. Therefore, for example, from the DC component to the high-frequency component, in the oblique direction as shown in FIG. 3A, and in the vertical direction as shown in FIG.
By selecting one of the three encoding order methods performed in the horizontal direction as shown in (1), the image quality is slightly inferior, but the encoding order determination process can be simplified and the processing speed can be increased.

When the code length of each coefficient is controlled so that the entire block has a predetermined code length, FIG.
When the input code "9" exceeds the above-mentioned predetermined length, the input code is deleted and the position of the input code is meaningless, so that the fixed-length encoding process can be simplified. When the input code "9" exceeds the predetermined length as shown in FIG. 5, the input code "9" is converted into an approximate code "9a" so as to fit into the input code "9", as shown in FIG. The image quality can be prevented from deteriorating as compared with the case where the input code “9” is discarded. At this time, if the difference between the input code "9" and its approximate code "9a" is equal to or greater than a predetermined value, the difference is discarded as shown in FIG. 4; Approximate code "9a"
By adopting, it is possible to further prevent the image quality from deteriorating.

Here, the coding unit 304 sequentially converts the DCT coefficient values yuv in the order determined by the coding order determination unit 303.
Therefore, it is necessary to inform the decoding unit 104 of this order. Therefore, the code length control unit 305 also adds the coding order information to the head of the code string of one block as shown in FIG.

Next, the decoding unit 104 will be described with reference to FIG. First, the decoding order determination unit 801 determines the decoding order based on the coding order information added to the head of the code string of one block as shown in FIG. Decode the columns sequentially. In the subsequent transform coefficient reproducing unit 803, the decoding unit 802
Are sequentially decoded by the decoding order determination unit 801
Is rearranged into the original order based on the decoding order determined by the above, and the DCT coefficient value yuv is reproduced by interpolating the missing coefficients with an arbitrary value. Next, coefficient conversion means 80
4 to convert the DCT coefficient value yuv into image data xij of 8.times.8 pixels per block.
Thus, the image data xij is reproduced in the original order.

[0028]

As described above, according to the first aspect of the present invention, the encoding order is determined based on the coefficient values subjected to the two-dimensional discrete cosine transform so that the important coefficients are encoded first. In this case, coding is performed based on the coding order, and control is performed so that the code length of the coefficient value for one block becomes a predetermined length. And the image quality can be improved.

According to the second aspect of the present invention, one of a plurality of preset encoding orders is selected, so that the processing for determining the encoding order can be simplified.

According to the third aspect of the present invention, when the code length of the coefficient value exceeds a predetermined length, the code exceeding the predetermined length is deleted, so that the fixed length coding processing can be simplified.

According to the fourth aspect of the present invention, when the code length of the coefficient value exceeds a predetermined length, the code exceeding the predetermined length is converted so as to be within the predetermined length and stored in the block. It is possible to prevent the image from deteriorating due to the conversion process.

According to the fifth aspect of the present invention, when the code length of the coefficient value exceeds a predetermined length, the code exceeding the predetermined length is converted so as to be within the predetermined length, and the difference between the code before the conversion and the code after the conversion is converted. Is larger than a predetermined value, the code exceeding the predetermined value is discarded, and if the value is smaller than the predetermined value, the converted code is stored in the block. Therefore, it is possible to prevent image deterioration due to the fixed length coding processing.

According to the sixth aspect of the present invention, since the information on the encoding order is stored in the block, the decoding device can decode in the normal order.

According to the seventh aspect of the present invention, when an image is encoded and decoded by the fixed length method, the image quality can be improved without lowering the encoding efficiency.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a facsimile apparatus to which an embodiment of an encoding apparatus and a decoding apparatus according to the present invention is applied.

FIG. 2 is a block diagram illustrating an encoding unit of FIG. 1 in detail.

FIG. 3 is an explanatory diagram showing another example of the encoding order determined by the encoding order determination unit in FIG. 2;

FIG. 4 is an explanatory diagram showing code length control of a code length control unit in FIG. 2;

FIG. 5 is an explanatory diagram showing another example of code length control of the code length control unit in FIG. 2;

FIG. 6 is an explanatory diagram showing a format process of a code length control unit in FIG. 2;

FIG. 7 is a block diagram illustrating a decoding unit of FIG. 1 in detail;

FIG. 8 is a block diagram showing an encoding unit and a decoding unit of a conventional facsimile apparatus.

[Explanation of symbols]

 301 Block division unit 302 Image conversion unit 303 Encoding order determination unit 304 Sequential encoding unit 305 Code length control unit 801 Decoding order determination unit 802 Sequential decoding unit 803 Transform coefficient reproduction unit 804 Coefficient conversion unit 805 Image reproduction unit

Claims (7)

[Claims] An image dividing unit that divides multi-valued image data into predetermined rectangular blocks; and a two-dimensional discrete cosine transform that converts image data in the blocks divided by the image dividing unit into coefficient values. Image conversion means, coding order determination means for determining the coding order so as to perform coding first from important coefficients based on the coefficient values converted by the image conversion means, and conversion by the image conversion means. Coding means for coding the calculated coefficient values based on the coding order determined by the coding order determination means; and a code length of one block of coefficient values coded by the coding means is predetermined. A fixed-length block encoding device comprising: a code length control unit that controls the length to be longer. 2. The image processing apparatus according to claim 1, wherein the encoding order determining unit selects one of a plurality of preset encoding orders based on the coefficient value converted by the image converting unit. Fixed-length block encoding device. 3. The code length control unit according to claim 1, wherein, when the code length of the coefficient value encoded by the encoding unit exceeds the predetermined length, the code length control unit deletes the code exceeding the predetermined length. 3. The fixed-length block encoding device according to 2. 4. When the code length of the coefficient value encoded by the encoding means exceeds the predetermined length, the code length control means converts the code exceeding the predetermined length so as to be within the predetermined length. 3. The fixed-length block coding device according to claim 1, wherein the fixed-length block coding device is stored in the block. 5. When the code length of the coefficient value encoded by the encoding means exceeds the predetermined length, the code length control means converts the code exceeding the predetermined length so as to be within the predetermined length, 3. The fixed code according to claim 1, wherein when the difference between the code before conversion and the code after conversion is equal to or more than a predetermined value, the code exceeding the predetermined value is discarded, and when the difference is equal to or less than the predetermined value, the code after conversion is stored in the block. Long block coding device. 6. The apparatus according to claim 1, wherein said code length control means stores the information on the coding order determined by said coding order determining means in the block. Fixed-length block coding device. 7. A fixed-length block decoding device that decodes a code encoded by the fixed-length block encoding device according to claim 6, wherein the coefficient value encoded by the encoding unit is encoded by the code. Decoding means for decoding based on the encoding order determined by the coding order determining means; image inverse transform means for inversely transforming the coefficient values decoded by the decoding means into image data for each block; A fixed-length block decoding device, comprising: image reproducing means for reproducing the image data for each block inversely transformed by the means into image data in the original order.