JP2001148852A - Image information converter and image information conversion method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the image quality of P and B pictures configured on the basis of an I picture by reducing the deterioration in the image quality accompanying the re-quantization of the I picture. SOLUTION: A quantization matrix changeover device 4 selects a prescribed quantization matrix used by an image information encoder 5 from a quantization matrix for an intra-macro-block into a quantization matrix for an inter-macro- block on the basis of attached information. The image information encoder 5 applies orthogonal transform to moving picture information supplied from an image information decoder 2 on the basis of the attached information and the inter-macro-block quantization matrix supplied from the quantization matrix changeover device 4 to encode the transformed moving picture information into the image compression information at a low bit rate.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image information conversion apparatus and an image information conversion method for converting the bit rate of compressed image information.

[0002]

2. Description of the Related Art In recent years, image information is handled as digital data, and the digital data is compressed by orthogonal transformation and motion compensation using the redundancy inherent in the image information, so that network media such as satellite broadcasting and cable television can be used. 2. Description of the Related Art Devices for transmitting data to a storage medium or recording on a storage medium such as an optical disk or a magnetic disk have been widely used. In such an apparatus, MPEG-2 (Moving Picture Experer) using discrete cosine transform as an image compression method is generally used.
ts Group phase-2) is used.

In recent years, standardization of digital television broadcasting using an image compression method such as MPEG-2 has been promoted. Digital television broadcasting standards include:
Standard resolution image (for example, if the number of effective lines in the vertical direction is 57
6), a standard corresponding to a high-resolution image (for example, the number of effective lines in the vertical direction is 1152), and the like.

Incidentally, the image information of this high-resolution image is enormous, and even if it is compressed using an encoding method such as MPEG-2, a large amount of code (bit rate) is required to obtain a sufficient image quality. Required. For example, in the case of a 30 Hz interlaced scan image having an image frame of 1920 × 1080 pixels, a code amount of about 18 to 22 Mbps or more is required.

[0005] Therefore, when transmitting such a high-resolution image to a network medium such as a satellite broadcast or a cable television, the code amount must be further reduced in accordance with the bandwidth of the transmission path. Similarly,
Even when such a high-resolution image is recorded on a storage medium such as an optical disk or a magnetic disk, the code amount must be further reduced according to the recording capacity of the medium. Further, it is conceivable that such a necessity of reducing the code amount occurs not only in a high-resolution image but also in a standard-resolution image (for example, a 30 Hz interlaced scan image having an image frame of 720 × 480 pixels).

Means for solving such a problem include hierarchical coding (scalability) and image information conversion (transcoding). In the MPEG-2, the SNR scalability is standardized for the former, and using this, the high-SNR image compression information (bit stream) and the low SNR image compression information (bit stream) are hierarchically encoded. . However, in order to perform hierarchical encoding, it is necessary that a predetermined value such as a bandwidth or a storage capacity is known at the time of encoding, but it is often unknown in an actual system. Therefore,
It can be said that the latter is a more flexible system in accordance with the actual system.

In a conventional image information conversion apparatus (transcoder) using the latter image information conversion (transcoding), a decoding section for decoding or partially decoding image compression information (bit stream) to be input is provided. And an encoding unit that re-encodes the output of the decoding unit is connected in parallel, and image information is supplied from the decoding unit to the encoding unit in two regions, a spatial domain and a frequency domain.

A conventional image information conversion apparatus in which image information is supplied from a decoding unit to an encoding unit in the former spatial domain,
Although the amount of calculation processing is large, it is possible to suppress the deterioration of the decoded image of the compressed image information (bit stream) to be output, and it is mainly used for applications such as broadcasting equipment. On the other hand, the conventional image information conversion apparatus in which image information is supplied from the decoding unit to the encoding unit in the latter frequency domain causes a slight deterioration in image quality as compared with the former image information conversion apparatus, but has a smaller effect. It can be realized with an arithmetic processing amount, and is mainly used for applications of consumer appliances.

Next, a conventional image information conversion apparatus used in each of the spatial domain and the frequency domain will be described with reference to the drawings.

First, a conventional image information conversion device used in the spatial domain will be described. FIG. 19 shows a conventional image information conversion device used in this spatial domain.

As shown in FIG. 19, a conventional image information conversion device 100 includes an image information decoding device 101, an additional information buffer 102, and an image information encoding device 103.

The conventional image information conversion apparatus 100 is an apparatus for reducing the amount of code of image compression information (bit stream), and supplies image information from an image information decoding apparatus 101 to an image information encoding apparatus 103. In the spatial domain.

First, in the conventional image information conversion apparatus 100, the image information decoding apparatus 101 receives high-bit-rate image compression information. This image information decoding device 101
Once completely decodes the high bit rate image compression information and outputs baseband video data. At the same time, the additional information buffer 102 is supplied from the image information decoding apparatus 101 with information (hereinafter referred to as additional information) used by the image information decoding apparatus 101 at the time of decoding. Store the information.

The additional information includes, for example, information for each macroblock such as a motion vector, a prediction mode, a DCT mode, a quantization scale code, a GOP header (Groupe of Picture Header), and a picture header (Pict).
ure Header), sequence header (Sequence Heade)
r), Sequence Display Extension
ion) Picture Coding Exte
nsion), Quantization matrix extension (Quantization M
atrix Extension), Picture Display Extension (PictureDisp)
lay Extension).

The image information coding apparatus 102 is provided with a target code amount (target bit rate) lower than the code amount (high bit rate) of the input image compression information in advance. , Based on the additional information acquired from the additional information buffer 102. That is, the image information encoding device 103 uses the image information decoding device 10 based on the target code amount and the additional information.
1 to re-encode the baseband video data obtained as an output, and output low bit rate image compression information. As described above, by using the additional information stored in the additional information buffer 102, the image information encoding device 103 can reduce an increase in the amount of arithmetic processing and image quality deterioration due to re-encoding.

For example, when image information is generally encoded, a large amount of calculation processing is required for motion vector search. However, in the conventional image information conversion apparatus 100, each macro stored in the additional information buffer 102 By using a motion vector and a prediction mode for each block, encoding processing can be performed without performing a motion vector search.

Next, a conventional image information conversion device used in the frequency domain will be described. FIG. 20 shows a conventional image information conversion device used in this frequency domain.

As shown in FIG. 20, a conventional image information converter 110 includes a code buffer 111, a compression information analyzer 112, a variable length decoder 113, an inverse quantizer 114, a band limiter 115 and the quantization device 11
6, the information buffer 117, and the variable length coding device 118
, A code buffer 119, and a code amount control device 120.

The code buffer 111 receives image compression information (bit stream) having a large code amount (high bit rate) and stores the input image compression information. In the code buffer 111, image compression information (bit stream) encoded so as to satisfy the VBV (Video Buffer Verifier) constraint defined in MPEG-2.
, The overflow and / or underflow does not occur. Then, the code buffer 111 supplies the stored image compression information to the compression information analysis device 112.

The compression information analysis device 112 is an MPEG-2
Based on the syntax (syntax) specified in (1), information necessary for each process described below (hereinafter, referred to as analysis result information) is extracted from the image compression information (bit stream) supplied from the code buffer 111, The extracted analysis result information is transmitted to the variable length decoding device 113 and the information buffer 117.
To supply. The compression information analysis device 112 includes, among the analysis result information, a code amount control device 12 described later.
0, which is necessary for processing in picture 0, and quantization scale information (q_sc) which is information on a quantization value for each macroblock.
ale) and the like are supplied to the information buffer 117.

The variable-length decoding device 113 variably converts data encoded as a difference value between adjacent blocks with respect to the DC component of the intra macroblock of the image compression information supplied from the compression information analysis device 112. By performing long decoding and performing variable length decoding on data encoded by the run and level for other coefficients, a quantized one-dimensional discrete cosine transform coefficient is obtained. Then, the variable-length decoding device 113 inversely transforms the one-dimensionally arranged discrete cosine transform coefficients based on information on the scanning method (zigzag scan or alternate scan) included in the analysis result information extracted by the compression information analysis device 112. Scan and rearrange into quantized two-dimensional discrete cosine transform coefficients. The variable length decoding device 113 supplies the two-dimensional array and the quantized discrete cosine transform coefficients to the inverse quantization device 114.

The inverse quantization device 114 calculates the quantization width and the quantization matrix included in the analysis result information,
Dequantize the two-dimensional array and the quantized discrete cosine transform coefficients. The inverse quantization device 114 supplies the inversely quantized discrete cosine transform coefficients to the band limiting device 115.

The band limiting device 115 is connected to the inverse quantization device 11
4 with respect to the discrete cosine transform coefficients supplied from
The band of the high frequency component coefficient in the horizontal direction is limited for each T block. Then, the band limiting device 115 supplies the discrete cosine transform coefficient subjected to the band limitation to the quantization device 116.

The quantizing device 116 includes a band limiting device 115
Transforms the supplied 8 × 8 discrete cosine transform coefficients into a quantization width controlled by the code amount control device 120 according to the target code amount (target bit rate) of the output image compression information (bit stream). Based on the quantization. Then, the quantization device 116 supplies the quantized discrete cosine transform coefficients to the variable length coding device 118.

The information buffer 117 includes, for example, picture coding type information (picture_coding_type) and quantization scale information (q_
Analysis result information such as scale) is stored. Then, the information buffer 117 supplies the stored analysis result information to the code amount control device 120.

The variable-length coding device 118 includes the quantization device 1
Variable-length coding of the quantized discrete cosine transform coefficient supplied from 16 is performed, and the variable-length-coded discrete cosine transform coefficient is supplied to the code buffer 119.

The code buffer 119 is a buffer memory for keeping the amount of output low-bit-rate image compression information constant, and receives a small amount of code (low bit-rate) image compression information (bit stream). The input image compression information is stored. This code buffer 1
19, VBV (VideoBuf) specified by MPEG-2
Since the image compression information (bit stream) encoded so as to satisfy the constraint condition of the “fer verifier” is stored, overflow and / or underflow does not occur. Then, the code buffer 119 outputs the stored image compression information and supplies it to the code amount control device 120.

The code amount control unit 120 controls a predetermined target so that the image compression information after the variable length coding by the variable length coding unit 118 does not cause overflow and / or underflow in the code buffer 119. Based on the code amount (target bit rate) and the analysis result information obtained from the information buffer 117, the quantization width of the quantization matrix used in the quantization device 116 is controlled.

In the image information conversion device 110 configured as described above, the inverse quantization device 114 converts the two-dimensional array and the quantized discrete cosine transform coefficients supplied from the variable length decoding device 113 into analysis results. The inverse quantization is performed based on the quantization width and the information related to the quantization matrix included in the information, and the inversely quantized discrete cosine transform coefficient is determined by the band limiting device 1.
15 Then, the quantization device 116 converts the 8 × 8 discrete cosine transform coefficient supplied from the inverse quantization device 114 via the band limiting device 115 into the code amount control device 12
The quantization is performed based on the quantization width controlled by 0. Then, the quantization device 116 supplies the quantized discrete cosine transform coefficients to the variable length coding device 118. By performing such processing, image compression information of a low bit rate is output from the code buffer 119.

[0030]

SUMMARY OF THE INVENTION By the way, CCIR
(International Radio Consultative Committee) Test sequence "Mobile &Calendar"
Luminance signal for an original image of a decoded image of compressed image information (bit stream) encoded by an MPEG-2 image information encoding device (hereinafter, referred to as an MPEG-2 image information encoding device) compliant with Test Model 5 FIG. 21 shows the transition of the signal-to-noise ratio (hereinafter, referred to as pSNR) for each frame.

Here, the coding conditions are as follows: the bit rate is 6 Mbps, the structure of the GOP (Group of Pictures) is N = 15, M = 3. Note that N is a GOP
Where M is an I picture or P
This is the cycle at which the picture appears.

At this time, if the mean square error with the original image for each frame is MSE, the pSNR is expressed by the following equation (1).

[0033]

(Equation 1)

In FIG. 21, for example, 3, 9, 15
An I picture composed of such frame numbers has a higher pSNR than a neighboring P picture or B picture. This is because, in an MPEG-2 image information encoding device, an I picture has a target code amount (target bit).
Is set higher than the P picture or the B picture. Therefore, when the image quality of an I picture is improved, the image quality of a P picture or a B picture formed by referring to the I picture is also improved.

On the other hand, the CCIR test sequence “Mob
ile & Calendar ”without performing code amount control,
Without considering buffer overflow and / or underflow, the quantization value is fixed to 1 and MPEG-2
FIG. 22 shows the transition of the pSNR of the luminance signal for each frame with respect to the original image of the decoded image of the image compression information (bit stream) encoded by the image information encoding device.

In FIG. 22, contrary to the case of FIG. 21, an I picture having a frame number of, for example, 3, 9, 15, etc., has a lower pSNR than a neighboring P picture or B picture. I have. That is, the I picture is
The image quality is lower than that of the neighboring P picture or B picture.

This is due to the quantization matrix used in the MPEG-2 image information encoding device.
That is, in the MPEG-2 image information encoding apparatus, a quantization matrix as shown in FIGS. 23A and 23B is defined by a default value for each of an intra macroblock and an inter macroblock. Therefore, the intra macroblock is quantized twice by the quantization matrix shown in FIG. Therefore, as shown in FIG. 22, the requantization distortion in the high frequency component of the I-picture is higher than that of the P-picture or the B-picture, even though the I-picture is assigned a larger code amount (bit). It is getting bigger.

It should be noted that MPEG-2, which is used in practice,
In the image information encoding device, the quantization matrix shown in FIG. 23C defined by Test Model 5 is generally used instead of the quantization matrix defined in FIG. The experimental results shown in FIGS. 21, 22, 24, and 25 are all used as quantization matrices for intra macroblocks and inter macroblocks, respectively, as shown in FIG.
3 (a) and FIG. 23 (c).

Further, the CCIR test sequence “Mob
ile & Calendar ”, image compression information (bit stream) compressed to 6 Mbps is input, and FIG.
Alternatively, the code amount (bit rate) is further reduced using the image information conversion apparatus shown in FIG.
FIGS. 24 and 25 show the transition of the pSNR of the luminance signal with respect to the original image of the decoded image of the image compression information (bit stream) output as for each frame.

The result shown in FIG. 24 is obtained without using the additional information buffer 102 in FIG.
01 and the image information encoding device 103 are independently operated, and the motion vector is recalculated.

The results shown in FIG. 25 show that the band limiting device 115 shown in FIG. 20 does not reduce high frequency components, and the motion compensation error is corrected by an 8 × 8 discrete cosine for both P and B pictures. This is performed for all components of the transform coefficient, and a capacity of 15 frames is secured as the capacity of the feedforward buffer shown in FIG. Then, the normalized activity N_act is given by the following equation (2).
It is represented by

[0042]

(Equation 2)

Here, the tendency of the image quality in FIGS. 24 and 25 is the same as that shown in FIG. 22 described above. For example, the image quality of the I picture composed of the frame numbers such as 18, 33, 48 is Is lower than that of the P picture or the B picture.

The reason for this is as described above with reference to FIG.
The same can be said for the experimental results shown in FIG. That is, by the operation of the code amount control device 120 described above, FIGS.
Also in the experiment shown in FIG. 5, the intra macroblock is quantized twice by the quantization matrix shown in FIG. Therefore, the requantization distortion in the high-frequency component of the I picture is larger than that of the P picture or the B picture, even though a larger amount of code (bits) is assigned to the I picture.

As described above, the negative effect of requantization distortion for intra macroblocks is relatively larger than the positive effect of allocating a larger amount of code (bits) to I-pictures. 24 and 25, the picture quality of the I picture is low. Subjectively, deterioration of the image quality in the I picture is observed as a flash phenomenon once every 15 frames (0.5 seconds). Further, such a situation is a factor that hinders the improvement of the image quality of the P picture and the B picture formed with reference to the I picture.

Accordingly, the present invention has been made in view of such circumstances, and by reducing image quality deterioration due to requantization in an I picture, a P picture and a P picture constructed based on the I picture are reduced. It is an object of the present invention to provide an image information conversion device and an image information conversion method for reducing image quality deterioration of a B picture.

[0047]

In order to achieve the above-mentioned object, an image information conversion apparatus according to the present invention compresses an image signal by performing an orthogonal transformation in units of an orthogonal transformation block composed of predetermined pixel blocks. An image information conversion apparatus for converting the first compressed image information of the first bit rate into second compressed image information of a second bit rate lower than the first bit rate; Decoding means for decoding the first image compression information to generate moving image information; and an intra macro block which is a quantization matrix for intra-frame encoding used when the first image compression information is generated. Matrix switching means for switching a quantization matrix for use with an inter-macroblock quantization matrix, which is a quantization matrix for inter-frame coding, and the decoding means comprising: Encoding means for orthogonally transforming the moving image information generated by the decoding means and encoding the moving image information into the second image compression information based on the additional information used when decoding the image compression information, The encoding unit quantizes the intra macroblock based on the quantization matrix for the inter macroblock switched by the quantization matrix switching unit, and the decoding unit decodes the first image compression information. The inter macro block is quantized based on the used quantization matrix for the inter macro block.

In this image information conversion device, the quantization matrix switching means switches the quantization matrix for the intra macro block to the quantization matrix for the inter macro block. The encoding unit quantizes the intra macroblock based on the inter-macroblock quantization matrix switched by the quantization matrix switching unit, and uses the quantization value when the decoding unit decodes the first image compression information. The inter macro block is quantized based on the quantization matrix for the inter macro block.

Further, the image information conversion device according to the present invention
A first image obtained by compression-encoding an image signal by performing orthogonal transformation in units of orthogonal transformation blocks each including a predetermined pixel block.
In the image information conversion device for converting the first image compression information of the second bit rate into the second image compression information of the second bit rate lower than the first bit rate, A first image input to the image compression information analysis unit based on analysis result information which is an analysis result analyzed by the image compression information analysis unit; An inverse quantization means for inversely quantizing an orthogonal transform coefficient of the compressed information; and an inverse quantization means based on a quantization width such that the output second image compression information has the second bit rate. Quantizing means for requantizing the orthogonally-transformed coefficient of the decompressed first image compression information, and quantization for intra-frame encoding used when the first image compression information is generated queue A quantization matrix switching means for switching a quantization matrix for an intra macroblock to a quantization matrix for an inter macroblock which is a quantization matrix for inter-frame coding, wherein the quantization means The intra macroblock is quantized based on the quantization matrix for the inter macroblock switched by the matrix switching unit and the quantization width, and the inverse quantization unit inversely quantizes the first image compression information. The inter macro block is quantized based on the quantization matrix for the inter macro block used at the time and the quantization width.

In this image information converter, the quantization matrix switching means switches the quantization matrix for the intra macroblock to the quantization matrix for the inter macroblock. The quantization means quantizes the intra macroblock based on the quantization matrix for the inter macroblock and the quantization width switched by the quantization matrix switching means, and the inverse quantization means performs the first image compression. The inter macro block is quantized based on the quantization matrix for the inter macro block and the quantization width used when the information is inversely quantized.

Further, in the image information conversion method according to the present invention, the first image compression at the first bit rate, in which the image signal is compression-coded by performing orthogonal transformation in units of orthogonal transformation blocks composed of predetermined pixel blocks. Information from the first
In the image information conversion method of converting into the second image compression information of the second bit rate lower than the bit rate of the first, the first image compression information is decoded to generate the moving image information, A quantization matrix for intra macroblocks, which is a quantization matrix for intra-frame encoding used when the first image compression information is generated, is replaced with a quantization matrix for inter-macroblocks, which is a quantization matrix for interframe encoding. , And quantizes the intra macroblock based on the additional information used when decoding the first image compression information and the switched quantization matrix for the inter macroblock. The inter macro block is quantized based on the additional information and the quantization matrix for the inter macro block used when decoding the first image compression information. And wherein the Rukoto.

In this image information conversion method, the quantization matrix for the intra macro block is switched to the quantization matrix for the inter macro block, and the additional information used when decoding the first image compression information is switched. The intra macro block is quantized based on the quantization matrix for the inter macro block, and the inter macro block is quantized based on the additional information and the quantization matrix for the inter macro block used when decoding the first image compression information. Quantize the macroblock.

Still further, in the image information conversion method according to the present invention, an image signal is compression-encoded by performing an orthogonal transformation in units of an orthogonal transformation block composed of a predetermined pixel block, and the first image of the first bit rate is encoded. In an image information conversion method for converting compression information into second image compression information having a second bit rate lower than the first bit rate, the first image compression information having the first bit rate may be used. Is input, the input first image compression information is analyzed, and the orthogonal transform coefficient of the input first image compression information is dequantized based on the analysis result information that is the analysis result. A quantization matrix for intra macroblocks, which is a quantization matrix for intra-frame encoding used when the first image compression information is generated, is replaced with a quantization matrix for inter-frame encoding. Switching the quantization matrix for inter macroblock are matrices, a quantization matrix for the inter-macro block is switched, the output is the second
The intra-macroblock is quantized based on the quantization width such that the image compression information of the second bit rate becomes the second bit rate, and the inter-macroblock used for inverse quantization of the first image compression information is quantized. The inter-macroblock is quantized based on the quantization matrix and the quantization width.

In this image information conversion method, the quantization matrix for the intra macro block is switched to the quantization matrix for the inter macro block, and the switched quantization matrix for the inter macro block and the second image compression output A quantization matrix for the inter macroblock used when the intra macroblock is quantized based on the quantization width at which the information becomes the second bit rate, and the first image compression information is inversely quantized. The inter macroblock is quantized based on the quantization width.

[0055]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a first embodiment to which the present invention is applied will be described with reference to the drawings.

The image information conversion apparatus according to the first embodiment to which the present invention is applied is, for example, an MPEG-2 (Moving Pic).
This is an apparatus that outputs a low bit rate image compression information by reducing the code amount (bit rate) of the image compression information (bit stream) encoded by the image expert group phase-2) method. In the image information conversion apparatus according to the first embodiment to which the present invention is applied, supply of the image information from the decoding unit that decodes the image information to the encoding unit that encodes the image information is performed in the spatial domain. Have been done. FIG. 1 shows an image information conversion apparatus according to a first embodiment to which the present invention is applied. MPEG-2 refers to a compression method of image information corresponding to both interlaced scan images and progressive scan images, and both standard resolution images and high resolution images.

As shown in FIG. 1, the image information conversion device 1 includes an image information decoding device 2, an additional information buffer 3,
The image processing apparatus includes a quantization matrix switching device 4 and an image information encoding device 5. The image information conversion device 1 is generally a device for reducing the code amount of image compression information (bit stream), and supplies image information from the image information decoding device 2 to the image information encoding device 5 in a spatial domain. Do with.

The image information decoding device 2 receives high-bit-rate image compression information, completely decodes the input high-bit-rate image compression information, and obtains a baseband video signal obtained as a result of the decoding. The data is supplied to the image information encoding device 5. Simultaneously with this processing, the image information decoding apparatus 2 supplies the additional information used for decoding the input high-bit-rate image compression information to the additional information buffer 3.

The additional information includes, for example, information for each macroblock such as a motion vector, a prediction mode, a DCT mode, a quantization scale code, a GOP header (Groupe of Picture Header), and a picture header (Pict).
ure Header), sequence header (Sequence Heade)
r), Sequence Display Extension
ion) Picture Coding Exte
nsion), Quantization matrix extension (Quantization M
atrix Extension), Picture Display Extension (PictureDisp)
lay Extension).

The additional information buffer 3 stores the additional information supplied from the image information decoding device 2. Specifically, the additional information buffer 3 is supplied from the image information decoding device 2 with information on two quantization matrices for the intra macroblock and the inter macroblock used by the image information decoding device 2 and stores the information. I do. That is, here
The additional information includes information on a quantization matrix for an intra macroblock as shown in FIG. 2A and information on a quantization matrix for an inter macroblock as shown in FIG. And

The additional information buffer 3 stores only the information on the inter-macroblock quantization matrix among the supplied information on the two quantization matrices,
In accordance with the control information from the quantization matrix switching device 4, the information is supplied to the quantization matrix switching device 4.

The quantization matrix switching device 4 uses the frame information used when the high bit rate image compression information input to the image information decoding device 2 is generated based on the additional information obtained from the additional information buffer 3. The quantization matrix for the intra macroblock which is the quantization matrix for the inner coding is switched to the quantization matrix for the inter macroblock which is the quantization matrix for the interframe coding.

More specifically, the quantization matrix switching device 4 selects only the information related to the quantization matrix for the inter macro block from the additional information stored in the additional information buffer 3, and transfers the selected information to the relevant information. Obtained from the additional information buffer 3. The quantization matrix switching device 4 is used when the high bit rate image compression information input to the image information decoding device 2 is generated based on the acquired information on the inter-macroblock quantization matrix. The quantization matrix for the intra macroblock is switched to the quantization matrix for the inter macroblock. After that, the quantization matrix switching device 4 supplies the switched quantization matrix for the inter macro block to the image information encoding device 5.

However, when the (0, 0) -th component of the switched inter-macro block quantization matrix is not 8, the quantization matrix switching device 4 performs, for example, the (0, 0) , 0) is converted to 8 to generate a quantization matrix, and the generated quantization matrix is supplied to the image information encoding device 5. This is because the MPEG-2 standard specifies that the (0, 0) -th component of the quantization matrix is 8.

The image information encoding device 5 is provided with a target code amount (target bit rate) lower than the code amount (high bit rate) of the input image compression information in advance. An encoding process is performed based on the additional information acquired from the information buffer 3 and the quantization matrix supplied from the quantization matrix switching device 4. That is, the image information encoding device 5 re-encodes the baseband video data supplied from the image information decoding device 2 based on the target code amount, the additional information, and the quantization matrix, and performs image compression at a low bit rate. Output information.

For example, the image information encoding device 5 quantizes the intra macro block based on the target code amount, the additional information, and the quantization matrix for the inter macro block switched by the quantization matrix switching device 4. I do. Further, the image information encoding device 5 includes the target code amount, the additional information, and the quantization matrix for the inter macro block used when the image information decoding device 2 decodes the input high bit rate image compression information. Quantizes the inter macroblock based on

In the image information conversion device 1 configured as described above, the high bit rate image compression information input to the image information decoding device 2 is once completely decoded by the image information decoding device 2, and The data is supplied to the image information encoding device 5 as video data of the band. Then, the baseband video data supplied to the image information encoding device 5 is re-encoded by the image information encoding device 5 based on the target code amount, the additional information, and the quantization matrix. Is output as image compression information.

Next, FIG. 4 shows a measurement result obtained by using the quantization matrix switching device 4 in the image information conversion device 1 according to the first embodiment to which the present invention is applied. FIG. 5 shows a measurement result in which the degree of image quality is improved by the measurement under the same conditions as the measured measurement using the pSNR of the luminance signal.

In this case, the input image compression information (bit stream) is shown in FIGS. 6A and 6C in the respective quantization matrices for the intra macroblock and the inter macroblock. Quantization matrix is used. Therefore, the quantization matrix for the intra macroblock used for the output image compression information (bit stream) is as shown in FIG. However, the quantization matrix for the inter macro block remains the quantization matrix shown in FIG.

Further, in the image information conversion apparatus 1, the change in the measurement result represented by the pSNR of the luminance signal when the quantization matrix is switched and when the quantization matrix is not switched is shown in FIG. Shown in FIG.

As shown in FIG. 5, in the image information conversion apparatus 1, by switching the quantization matrix,
As for the picture, the image quality is greatly improved by about 1.0 to 3.0 dB, and the flash phenomenon observed in FIG. 4 is not observed even in the subjective evaluation. The image quality of P-pictures and B-pictures formed based on I-pictures has also been improved.

Next, FIG. 8 shows a diagram in which the code amount (bit) assigned to each frame of the output image compression information (bit stream) in the measurement shown in FIG. 5 is measured.

In the image information conversion device 1, as shown in FIG. 8, the quantization matrix used in the image information encoding device 5 is switched to a quantization matrix for an inter macroblock, so that the high frequency components are quantized coarsely. At the same time, more codes (bits) are allocated to the I picture, and less codes (bits) are allocated to the P picture, thereby improving the picture quality of the I picture. The image quality of the P picture and the B picture based on this is improved.

As described above, the first embodiment to which the present invention is applied
In the image information conversion device 1 according to the embodiment of the present invention, the quantization matrix used in the image information encoding device 5 is higher in frequency band than the quantization matrix for the intra macroblock from the quantization matrix for the intra macroblock. By switching to a quantization matrix for an inter macroblock in which components are not coarsely quantized, image quality degradation in an I picture is prevented, and a flash phenomenon of an image is also subjectively avoided. The image quality of the composed P picture and B picture can also be improved.

Further, in the image information conversion apparatus 1 according to the first embodiment to which the present invention is applied, the quantization matrix for the inter macroblock is converted to both the intra macroblock and the inter macroblock. By using this, the quantization matrix switching device 4 does not need to include a storage medium and store the quantization matrix for switching.

In the image information conversion device 1 described above,
MPEG-2 image compression information (bit stream)
Is input, but if it is image compression information (bit stream) encoded by orthogonal transformation and motion compensation,
For example, MPEG-1 or H.264. 263 or the like may be input.

Next, a second embodiment of the present invention will be described with reference to the drawings.

The image information conversion apparatus according to the second embodiment to which the present invention is applied is also encoded by, for example, the MPEG-2 system, similarly to the image information conversion apparatus 1 according to the first embodiment. This is a device that outputs a low bit rate image compression information by reducing the code amount (bit rate) of the compressed image information (bit stream). In the image information conversion apparatus according to the second embodiment to which the present invention is applied, the supply of the image information from the decoding unit that decodes the image information to the encoding unit that encodes the image information is performed in the frequency domain. Have been done. FIG. 9 shows an image information conversion apparatus according to a second embodiment of the present invention.

As shown in FIG. 9, the image information converter 10 includes a code buffer 11 and a compression information analyzer 12.
, Variable-length decoding device 13, inverse quantization device 14, band limiting device 15, information buffer 16, quantization matrix switching device 17, quantization device 18, variable-length coding device 1
9, a code buffer 20, and a code amount control device 21.

The code buffer 11 receives image compression information (bit stream) of a large code amount (high bit rate) and stores the input image compression information. In this code buffer 11, VB defined by MPEG-2 is used.
Since image compression information (bit stream) encoded so as to satisfy the constraint condition of V (Video Buffer Verifier) is stored, overflow and / or underflow does not occur. And the code buffer 11
Transmits the stored image compression information to the compression information analysis device 12.
To supply.

The compression information analysis device 12 extracts information necessary for each process described later from the image compression information (bit stream) supplied from the code buffer 11 based on the syntax (syntax) specified by MPEG-2. And supplies the extracted information (hereinafter referred to as analysis result information) to the variable length decoding device 13 and the information buffer 16. The compression information analysis device 12 includes, among the analysis result information, picture encoding type information (picture_codi
ng_type) and quantization scale information (q_scale), which is information on the quantization value for each macroblock, are supplied to the information buffer 16.

The variable length decoding device 13 variably converts the data encoded as a difference value between the adjacent block and the DC component of the intra macroblock of the image compression information supplied from the compression information analysis device 12. By performing long decoding and performing variable length decoding on data encoded by the run and level for other coefficients, a quantized one-dimensional discrete cosine transform coefficient is obtained. Then, the variable-length decoding device 13 outputs information on a scanning method (a zigzag scan shown in FIG. 10A or an alternate scan shown in FIG. 10B) included in the analysis result information extracted by the compression information analysis device 12. , The inversely scanned one-dimensionally arranged discrete cosine transform coefficients are rearranged into quantized two-dimensional discrete cosine transform coefficients. The variable length decoding device 13 supplies the two-dimensional array and the quantized discrete cosine transform coefficients to the inverse quantization device 14.

The inverse quantization device 14 inversely quantizes the two-dimensional array and the quantized discrete cosine transform coefficients based on the information on the quantization width and the quantization matrix included in the analysis result information. The inverse quantization device 14 supplies the inversely quantized discrete cosine transform coefficients to the band limiting device 15.

The band limiting unit 15 limits the band of the horizontal high-frequency component coefficient for each DCT block with respect to the discrete cosine transform coefficient supplied from the inverse quantization unit 14.

FIG. 11 shows an example of the band limiting process of the high frequency component in the horizontal direction in the band limiting device 15. For example,
The band limiting device 15 performs the processing shown in FIG.
As shown in (a), of the 8 × 8 discrete cosine transform coefficients, only the values of 8 × 6 coefficients, which are low-frequency components in the horizontal direction, are stored, and the rest are replaced with 0. Also, the band limiting device 15
With respect to the color difference signal, as shown in FIG.
Of the 8 × 8 discrete cosine transform coefficients, only the values of the 8 × 4 coefficients, which are low-frequency components in the horizontal direction, are stored, and the rest are replaced with 0. By limiting the band of the high-frequency component of the discrete cosine transform coefficient in this way, it is possible to reduce the code amount (bit rate) in the frequency domain.

When the input image compression information (bit stream) is that of an interlaced scan image, the information on the time difference between fields is included in the vertical high-frequency component of the discrete cosine transform coefficient. Become. Therefore, limiting the band of the discrete cosine transform coefficient in the vertical direction leads to a significant deterioration in image quality. Therefore, the band limiting device 15 does not perform band limiting in the vertical direction.

Further, the band limiting device 15 limits the band of a color difference signal which is more difficult to be seen by humans than a luminance signal which is more easily seen by humans. As a result, this band limiting device 1
In No. 5, distortion of requantization can be reduced while minimizing image quality deterioration. When the amount of code (bit rate) to be reduced is small or when there is a circuit limitation, the band limitation of the luminance signal and the color difference signal may be the same.

Furthermore, the band limiting process of the horizontal discrete cosine transform coefficient in the band limiting device 15 is not limited to the process of setting the coefficient to 0 as shown in FIG.
For example, instead of replacing with 0, it is possible to similarly reduce the code amount (bit rate) by multiplying a previously prepared weight coefficient by the horizontal high-frequency component of the discrete cosine transform.

The band limiting device 15 converts the discrete cosine transform coefficient subjected to the band limiting as described above into a quantizing device 18.
To supply.

The information buffer 16 stores the compressed information analyzer 1
2, for example, picture coding type information (picture_coding_type) and quantization scale information (q_sc
ale) and other analysis result information are stored. Then, the information buffer 16 supplies the stored analysis result information to the quantization matrix switching device 17 and the code amount control device 21.

The quantization matrix switching device 17 uses the intra-macro block used when the high bit rate image compression information input to the code buffer 11 is generated based on the analysis result information obtained from the information buffer 16. Is switched to the quantization matrix for the inter macroblock.

More specifically, the quantization matrix switching device 17
From the additional information stored in the information buffer 16, only information regarding the quantization matrix for the inter macroblock is selected, and the selected information is obtained from the information buffer 16. The quantization matrix switching device 17 is used when the high bit rate image compression information input to the code buffer 11 is generated based on the acquired information on the inter-macroblock quantization matrix. The quantization matrix for the intra macro block is switched to the quantization matrix for the inter macro block. Thereafter, the quantization matrix switching device 17 supplies the switched quantization matrix for the inter macroblock to the quantization device 18.

However, when the (0,0) -th component of the switched inter-macro block quantization matrix is not 8, the quantization matrix switching device 17 performs, for example, the (0,0) -th component as shown in FIG. , 0) is converted to 8 to generate a quantization matrix, and the generated quantization matrix is supplied to the quantization device 18. This is also because the MPEG-2 standard specifies that the (0,0) th component of the quantization matrix is 8.

The quantization device 18 converts the 8 × 8 discrete cosine transform coefficient supplied from the band limiting device 15 into a quantization matrix supplied from the quantization matrix switching device 17 and a code amount control as described below. Quantization is performed based on a quantization width controlled by the device 21 and corresponding to a target code amount (target bit rate) of the output image compression information (bit stream). Then, the quantization device 18 supplies the quantized discrete cosine transform coefficient to the variable length coding device 19.

For example, the quantization device 18 quantizes an intra macroblock based on the quantization matrix for the inter macroblock switched by the quantization matrix switching device 17 and the quantization width. Also, the quantization device 18
Quantizes the inter macro block based on the quantization matrix for the inter macro block used when the inverse quantization device 14 inversely quantizes the input high bit rate image compression information and the quantization width. Become

The variable length coding device 19 is
Performs variable-length coding of the quantized discrete cosine transform coefficients supplied from the CDMA-PCB, and supplies the variable-length-coded discrete cosine transform coefficients to the code buffer 20.

The code buffer 20 is a buffer memory for keeping the amount of output low-bit-rate image compression information constant, and receives a small amount of code (low bit-rate) image compression information (bit stream). The input image compression information is stored. This code buffer 20
In VBV (Video Buffer) specified by MPEG-2
Since the image compression information (bit stream) encoded so as to satisfy the constraint condition of the verifier is stored, overflow and / or underflow does not occur. Then, the code buffer 20 outputs the stored image compression information and outputs the code amount control device 2.
Feed to 1.

[0098] The code amount control device 21 controls a predetermined target so that the image compression information after the variable length coding by the variable length coding device 19 does not cause overflow and / or underflow in the code buffer 20. Based on the code amount (target bit rate) and the analysis result information obtained from the information buffer 16, the quantization device 18
Controls the quantization width of the quantization matrix used in.

In the image information conversion device 1 configured as described above, the inverse quantization device 14 analyzes the two-dimensional array and the quantized discrete cosine transform coefficients supplied from the variable length decoding device 13, The inverse quantization is performed based on the information on the quantization width and the quantization matrix included in the information, and the inversely quantized discrete cosine transform coefficient is supplied to the band limiting device 15. Then, the quantization device 18 converts the 8 × 8 discrete cosine transform coefficient supplied from the inverse quantization device 14 via the band limiting device 15 into the quantization matrix supplied from the quantization matrix switching device 17 and the code amount. The quantization is performed based on the quantization width controlled by the control device 21. Then, the quantization device 18 supplies the quantized discrete cosine transform coefficient to the variable length coding device 19. By performing such processing, image compression information of a low bit rate is output from the code buffer 20.

Next, the processing in the code amount control device 21 will be described in detail.

MPEG-2 Test Mode applied to an image information encoding apparatus compatible with MPEG-2
l 5 (ISO / IEC JTC1 / SC29 / WG1
1N0400), first, GO
Each picture constituting P (I picture, P picture, B picture
The allocated bit amount for a picture) is allocated based on the bit amount allocated to a picture that has not been encoded in the GOP, including the picture to be allocated. Next, in order to make the allocated bit amount for each of the allocated pictures coincide with the actual code amount, the quantization scale code is calculated based on the storage capacities of three types of virtual buffers independently set for each picture. It is obtained by feedback control in units of macro blocks. Next, each macro is quantized so that the obtained quantized scale code is finely quantized in a flat portion where visual deterioration is conspicuous, and coarsely quantized in a complicated portion of a pattern in which deterioration is relatively inconspicuous. Varies depending on the activity of each block.

As described above, the image information conversion apparatus 10 according to the embodiment to which the present invention is applied also includes the Test Mod.
The code amount control is performed by an algorithm according to the method defined in el5.

However, if this technique is applied to the encoding section of the image information conversion device 10 shown in FIG. 9 as it is, the following two problems occur.

First, the first problem is that the MPEG-
2 This is a problem related to the content to be processed first in the method used in Test Model 5. That is, in the image information conversion apparatus compatible with MPEG-2, a GOP structure is provided in advance, and the first processing can be performed based on the GOP structure, whereas the image information conversion apparatus shown in FIG. In 10, the structure of the GOP is known by analyzing the syntax of all the information of one GOP in the input image compression information (bit stream). The length of this GOP is not always constant, and the MPEG-2 compliant image information converter detects a scene change, and adaptively determines the length of the GOP in the image compression information (bit stream). There are also things that are controlled by.

The second problem is that the above-described MPEG-
2 This is a problem related to the content to be processed last in the method used in Test Model 5. That is, in the MPEG-2 compliant image information conversion apparatus, the activity is calculated using the luminance signal pixel value of the original image. However, the image information conversion device 1 shown in FIG.
In the case of 0, since image compression information (bit stream) compatible with MPEG-2 is input, it is impossible to know the luminance signal pixel value of the original image.

Therefore, as a method of solving the first problem, there is a method of defining a pseudo GOP as described below and controlling the code amount based on the pseudo GOP. Here, the pseudo GOP refers to a pseudo GOP including one I picture and a plurality of P pictures and B pictures. The length of this pseudo GOP is variable and depends on how the I picture is detected in the image compression information (bit stream).

Hereinafter, a flow of a series of processing in the code amount control device 21 including a method for solving the first problem and the second problem will be described with reference to a flowchart shown in FIG.

First, in step S1 of FIG. 12, the information buffer 16 stores the picture_codin
It has a circular buffer for storing g_type. This circular buffer has 256 picturs equal to the maximum number of frames that can be included in one GOP specified by MPEG.
It has a storage capacity enough to store e_coding_type.
Each element of the circular buffer stores an initial value in advance.

Here, the information of each frame included in the image compression information (bit stream) is P, B, B, I,
Consider a case in which processing is performed up to B and B, and processing of the next P picture is performed. In this case, the image information conversion apparatus 10 first uses a feedforward buffer provided in the compression information analysis apparatus 12 to store several frames of picture data.
The _coding_type is read ahead and the elements of the circular buffer are updated. The size of the feedforward buffer is arbitrary, but is 6 frames in the circular buffer shown in FIG. The length of the pseudo GOP is set by referring to the pointer a indicating the current I picture and the pointer b indicating the next I picture from the state of the circular buffer shown in FIG. Further, the configuration of the pseudo GOP is set from the pointer d indicating the last frame of the feedforward buffer and the length of the pseudo GOP already set.

As described above, the structure of the pseudo GOP is set by the preparsing.

Subsequently, in step S2, the configuration of the pseudo GOP set as described above is changed to [B ₁ ,
B ₂ , P ₁ , B ₃ , B ₄ , I ₁ , B ₅ , B ₆ ,..., P _L , B
_M−1 , B _M ], the size of the pseudo GOP is L_
pgop is represented by the following equation (3).

[0112]

(Equation 3)

At this time, the target code amounts (target bits) T _i , T _p , and T _b of each picture (each frame) of the I picture, the P picture, and the B picture are represented by the following equations (4) and (5), respectively. ), Calculated by equation (6).

[0114]

(Equation 4)

[0115]

(Equation 5)

[0116]

(Equation 6)

Here, R is the amount of bits to be allocated to a picture that has not been coded in the GOP, including the picture to be allocated, and Θ is a frame that has already been processed in the pseudo GOP, and Ω is Assuming that a frame to be processed in the pseudo GOP, F is a frame rate, and B is a code amount (bit rate) of image compression information to be output, the following equations (7) and (8) are used. You.

[0118]

(Equation 7)

[0119]

(Equation 8)

X () is a parameter (global complexity measure) representing the complexity of each frame, and is the total code amount (bit number) of the frame when the compression information analyzer 12 performs pre-parsing. If S and Q, which is the average quantization scale code, are calculated in advance, they are expressed by the following equation (9).

[0121]

(Equation 9)

Further, K _p and K _b are MPE
G-2 I specified in Test Model 5
It is a ratio of the quantized scale code of the P picture and the B picture based on the quantized scale code of the picture, and is represented by the following equation (10).

[0123]

(Equation 10)

[0124] Then, it is assumed that K _p and K _b is at the value represented by the above formula (10), always the overall image quality is optimized.

Subsequently, in step S3, in order to match the actual generated code amount with the allocated bit amount (T _i , T _p , T _b ) for each picture calculated in step 2, it is set independently for each picture type. Based on the capacities of the three types of virtual buffers, the quantization scale code is obtained by feedback control in units of macroblocks.

First, prior to the j-th macroblock encoding, the occupancy of the virtual buffer is expressed by the following equations (11), (12), and (13).

[0127]

[Equation 11]

[0128]

(Equation 12)

[0129]

(Equation 13)

However, “d ₀ ⁱ ”, “d ₀ ^p ”, and “d ₀ ^b ” shown in the equations (11) to (13) are the initial values of the virtual buffers of the I, P, and B pictures. Occupancy
“B _j ” is the amount of generated bits from the beginning of the picture to the j-th macroblock, and “MB_cnt” is the number of macroblocks in one picture. Virtual buffer occupancy at the end of picture encoding (d _{MB_cnt} ⁱ , d _{MB_cnt} ^p , d
_{MB_cnt} ^b ) are the same picture type, and the initial values (d ₀ ⁱ ,
d ₀ ^p , d ₀ ^b ).

Next, the quantization scale code for the j-th macroblock is represented by the following equation (14).

[0132]

[Equation 14]

However, “r” shown in the equation (14) is a variable called a reaction parameter for controlling the response of the feedback loop, and is given by the following equation (15).

[0134]

(Equation 15)

The initial value of the virtual buffer at the start of encoding is given by the following equation (16).

[0136]

(Equation 16)

Subsequently, in step S4, the quantization scale Q of each macroblock in the input image compression information (bit stream) is calculated by using the luminance signal pixel value of the original image at the time of encoding. It is. Therefore, first, when preparsing is performed, the compression information analysis device 12 extracts a quantization scale Q and a code amount (number of bits) B of each macroblock in the frame, and the extracted quantization The scale Q and the code amount (the number of bits) B are stored in the information buffer 16. At the same time, the compression information analyzer 12 calculates in advance the average values E (Q) and E (B) of Q and B of the entire frame or the average value E (QB) of the product thereof, and calculates these values. Is stored in the information buffer 16.

In the code amount control device 21, the normalized activity N_act is calculated based on the information on Q and B stored in the information buffer 16 by the following equations (17) and (1).
8), and is represented by any one of equations (19).

[0139]

[Equation 17]

[0140]

(Equation 18)

[0141]

[Equation 19]

Of these, equations (18) and (19) are equivalent processing. Thus, adaptive quantization is performed based on the normalized activity N_act calculated in the DCT domain. Then, the image quality is determined by the signal-to-noise ratio (pSNR).
When the evaluation is made by using the expression (17), the image quality is higher in the expression (17), but the subjective image quality is better represented by the expression (18) or the expression (19).

Subsequently, in step S5, first, the quantization value in the input image compression information (bit stream) for a predetermined macroblock is represented by Q1, and the output amount represented by the above-described method in the code amount control device 21 is obtained. The quantization value for the compressed image information (bit stream) to be performed is Q2. Then, since the image information conversion apparatus 10 is for reducing the code amount (bit rate), when Q1> Q2, the macroblock once coarsely quantized is obtained from the result of requantization. This means that it has been finely quantized. Distortion due to coarse quantization is not reduced by fine requantization. Further, since many bits are used for this macroblock, the number of bits allocated to other macroblocks is reduced, and the image quality is further degraded. Therefore, when Q1> Q2, Q1 = Q2.

That is, if Q1> Q2, Q1 is output, while if Q1> Q2, Q2 is output.

The discrete cosine transform coefficients requantized through the above processing are supplied from the quantization device 18 to the variable length coding device 19.

The variable length coding device 19 includes a quantization device 18
From the quantized discrete cosine transform coefficients supplied from
Encode so that the average code length becomes short. At that time, the variable-length encoding device 19 encodes the difference with the DC component coefficient of one block before as a prediction value for the DC component of the discrete cosine transform coefficient, and scans a predetermined scanning method for the other components. After rearrangement into one-dimensional array data based on (zigzag scan or alternate scan), variable-length encoding is performed using a pair of the number of consecutive 0 coefficients (run) and the non-zero coefficient (level) as an event.

[0147] Then, when scanning is performed within the DCT block, the quantizer 18 outputs a code called EOB (End Of Block) if all the coefficients thereafter become 0. Then, the variable length coding for the block is completed.

The variable-length coding device 19 arranges discrete cosine transform coefficients into one-dimensional data by the alternate scan method regardless of the scan method of the inputted high-code-rate (high bit rate) image compression information. You may. The reason why the discrete cosine transform coefficients are arranged in one-dimensional data by the alternate scan method is as follows.

That is, it is assumed that the discrete cosine transform coefficient of a predetermined block of the input image compression information (bit stream) is, for example, as shown in FIG. In FIG. 14, coefficients indicated by ● are non-zero coefficients,
The coefficient indicated by ○ is a 0 coefficient. Assuming that the horizontal high-frequency component of the discrete cosine transform coefficient is 0 for such a discrete cosine transform coefficient, the distribution of non-zero coefficients is, for example, as shown in FIG.
The result is as shown in FIG. When the discrete cosine transform coefficient in which the horizontal high frequency component shown in FIG. 14B is set to 0 is re-encoded by zigzag scan, the scan number of the last non-zero coefficient becomes 50 (see FIG. 10A). On the other hand, when the scan conversion is performed and the encoding is performed again by the alternate scan, the scan number of the last non-zero coefficient becomes 44 (see FIG. 10B). For this reason, when performing variable-length coding on the discrete cosine transform coefficient with the horizontal high-frequency component set to 0, scanning by the alternate scan method sets the EOB signal with a scan number earlier than that in the zigzag scan. can do. Therefore, a finer value can be assigned as the quantization width, and quantization distortion accompanying re-quantization can be reduced.

The discrete cosine transform coefficients subjected to variable length coding by the variable length coding device 19 are stored in the code buffer 2.
0, temporarily stored in the code buffer 20, and then output as compressed image information in a bit stream structure defined by MPEG-2.

Next, the quantization matrix switching device 1 in the image information conversion device 10 according to the second embodiment of the present invention.
FIG. 15 shows the measurement results obtained by using FIG. 7, and FIG. 16 shows the measurement results in which the image quality is improved by the measurement under the same conditions as the measurement used at this time by the pSNR of the luminance signal. Shown in

In this case, the input image compression information (bit stream) is shown in FIG. 6A and FIG. 6C in the respective quantization matrices for the intra macroblock and the inter macroblock. Quantization matrix is used. Therefore, the quantization matrix for the intra macroblock used for the output image compression information (bit stream) is as shown in FIG. However, the quantization matrix for the inter macro block remains the quantization matrix shown in FIG.

Further, in the image information conversion device 10, the change in the measurement result represented by the pSNR of the luminance signal when the quantization matrix is switched and when the quantization matrix is not switched is shown in FIG. , Shown in FIG.

As shown in FIG. 16, in the image information conversion apparatus 10, by switching the quantization matrix, the image quality of the I picture is improved by about 0.4 dB. The observed flash phenomenon is no longer observed. The image quality of P-pictures and B-pictures formed based on I-pictures has also been improved.

Next, FIG. 17 shows a diagram in which the amount of code (bits) assigned to each frame of the output image compression information (bit stream) in the measurement shown in FIG. 16 is measured.

In the image information conversion apparatus 10, as shown in FIG. 17, the quantization matrix used in the quantization
By switching to the inter-macroblock quantization matrix, coarse quantization of high-frequency components is prevented.

As described above, the second embodiment to which the present invention is applied
In the image information conversion device 10 according to the embodiment of the present invention, the amount of code (bit rate) can be reduced by transferring data of each block in the frequency domain. Compared with the conventional image information conversion apparatus, the amount of calculation is reduced, and the circuit configuration can be significantly reduced.

Further, in the image information conversion apparatus 10 according to the second embodiment to which the present invention is applied, the quantization matrix used in the quantization apparatus 18 is converted from the quantization matrix for the intra macroblock by using this intra macroblock. By switching to a quantization matrix for inter-macroblocks that does not coarsely quantize high-frequency components as compared to a quantization matrix for image quality degradation in I-pictures is prevented, and the image flash phenomenon is also subjectively avoided. Accordingly, the image quality of the P picture and the B picture formed based on the I picture can be improved.

Further, in the image information conversion apparatus 10 according to the second embodiment to which the present invention is applied, the quantization matrix for the inter macroblock is converted to both the intra macroblock and the inter macroblock. By using, the quantization matrix switching device 17 includes a storage medium,
There is no need to store a quantization matrix for switching.

In the above-described image information conversion apparatus 10, image compression information (bit stream) according to MPEG-2 is input, but image compression information (bit stream) encoded by orthogonal transformation and motion compensation is used. If there is, for example, MPEG-1 or H.264. 263 or the like may be input.

Next, a third embodiment to which the present invention is applied will be described with reference to the drawings.

The image information conversion apparatus according to the third embodiment to which the present invention is applied is coded by, for example, the MPEG-2 method, similarly to the image information conversion apparatus 1 according to the first embodiment. This is a device that outputs a low bit rate image compression information by reducing the code amount (bit rate) of the compressed image information (bit stream). In the image information conversion apparatus according to the second embodiment to which the present invention is applied, the supply of the image information from the decoding unit that decodes the image information to the encoding unit that encodes the image information is performed in the frequency domain. Have been done. FIG. 18 shows an image information conversion apparatus according to a third embodiment to which the present invention is applied.

In describing the image information conversion apparatus 30, the same components as those of the image information conversion apparatus 10 according to the second embodiment are given the same reference numerals in the drawings, and detailed descriptions thereof will be given. Description is omitted.

As shown in FIG. 18, the quantization matrix switching device 17 generates the high bit rate image compression information input to the code buffer 11 based on the analysis result information obtained from the information buffer 16. Is switched to the quantization matrix for the intra macroblock used in the above.

More specifically, the quantization matrix switching device 17
From the additional information stored in the information buffer 16, only information regarding the quantization matrix for the inter macroblock is selected, and the selected information is obtained from the information buffer 16. The quantization matrix switching device 17 is used when the high bit rate image compression information input to the code buffer 11 is generated based on the acquired information on the inter-macroblock quantization matrix. The quantization matrix for the intra macro block is switched to the quantization matrix for the inter macro block. Thereafter, the quantization matrix switching device 17 supplies the switched quantization matrix for the inter macroblock to the quantization device 18.

However, when the (0,0) -th component of the switched inter-macroblock quantization matrix is not 8, the quantization matrix switching device 17 outputs the (0, 0) component as shown in FIG. , 0) is converted to 8 to generate a quantization matrix, and the generated quantization matrix is supplied to the quantization device 18. This is also because the MPEG-2 standard specifies that the (0,0) th component of the quantization matrix is 8.

The image information conversion device 30 includes a code buffer 1
1, a compressed information analyzer 12, and a variable length decoder 13
, Inverse quantization device 14, adder 40, band limiting device 15, information buffer 16, quantization matrix switching device 17
, A quantization device 18, a variable length coding device 19, a code buffer 20, a code amount control device 21, and a motion compensation error correction device 50.

The adder 40 is provided between the inverse quantization device 14 and the band limiting device 15. The adder 40 subtracts the motion compensation error correction coefficient generated by the motion compensation error correction device 50 from the discrete cosine transform coefficient obtained by inverse quantization by the inverse quantization device 14.

The motion compensation error correction device 50 corrects the motion compensation error generated when the quantized device 18 requantizes the discrete cosine transform coefficient inversely quantized by the inverse quantizer 14. Generate

Next, the cause of the occurrence of the motion compensation error will be described.

First, assume that the pixel value of the original image is O, and the quantization width of the input image compression information (bit stream) having a high code amount (high bit rate) with respect to the pixel value O of the original image is Q ₁ . the quantization width for the pixel value O in the original image of the image compression information of a low code amount after re-encoding (low bit rate) (bit stream) and Q _2. Then, let the pixel values of the reference image decoded with these quantization widths Q ₁ and Q ₂ be L (Q ₁ ) and L (Q ₂ ), respectively.

At the time of encoding, the pixels of the inter macro block are, for example, encoded by the image information conversion device 3 shown in FIG.
A difference value “OL (Q ₁ )” is calculated by the adder 40 of 0, and a discrete cosine transform is applied to the difference value “OL (Q ₁ )”. At the time of decoding, the pixels of the inter macroblock encoded in this way have a difference value “OL”.
(Q ₁ ) ”is subjected to an inverse discrete cosine transform, and a reference image“ L (Q ₁ ) ”generated by motion compensation is subtracted from the difference value“ OL (Q ₁ ) ”. The value O is decoded.

On the other hand, the pixels of the inter macroblock are
When the amount of code (bit rate) is reduced by the image information conversion device 10 shown in FIG. 9, the quantization width of the difference value “OL (Q ₁ )” is set to Q1 by the inverse quantization device 14 and the quantization device 18. To Q2. The pixels of the inter macro block whose code amount has been reduced in this way are decoded at the time of decoding, assuming that the difference value “OL (Q ₂ )” is coded with the quantization width Q _2. .

Here, since the amount of code is reduced by changing the quantization width in the image information converter 10, Q ₁ = Q ₂
Does not hold, and a quantization error occurs at the time of decoding the inter macroblock. Therefore, an error accompanying the motion compensation occurs in the P picture and the B picture encoded by the inter macro block.

The error generated in the P picture propagates to a P picture or a B picture which uses the P picture as a reference picture, which leads to further deterioration of image quality. As described above, the image quality deteriorates due to accumulation of errors due to the motion compensation of the GOP, and a phenomenon (drift) occurs that the next GOP returns to the good image quality again at the head.

In the motion compensation error correction device 50 of the image information conversion device 30 according to the third embodiment, a motion compensation error correction coefficient is generated, and the discrete cosine transform coefficient inversely quantized by the inverse quantization device 14 is used. Subtraction is performed to correct the above motion compensation error.

Subsequently, the motion compensation error correction device 50 will be described.

The motion compensation error correction device 50 includes an inverse quantization device 51, an adder 52, and an inverse discrete cosine transform device 53.
, A video memory 54, a motion compensation prediction device 55, and a discrete cosine transform device 56.

The inverse quantizer 51 inversely quantizes the discrete cosine transform coefficients requantized by the quantizer 18 based on the quantization matrix used in the quantizer 18. The discrete cosine transform coefficient inversely quantized by the inverse quantization device 51 is supplied to the adder 52.

The adder 52 subtracts the discrete cosine transform coefficient from which the motion compensation error correction coefficient has been subtracted by the adder 40 from the discrete cosine transform coefficient inversely quantized by the inverse quantizer 51, and performs inverse discrete cosine transform. It is supplied to the device 53.

The inverse discrete cosine transform unit 53 includes an adder 5
Inverse discrete cosine transform is performed on the discrete cosine transform coefficient supplied from step 2. The result obtained by performing the inverse discrete cosine transform is stored in the video memory 54 as motion compensation error correction information.

The motion compensation prediction device 55 includes motion compensation prediction mode information (field motion compensation prediction mode or frame motion compensation prediction mode, and image compression information (bit stream) having a high code amount (high bit rate) input. , Forward prediction mode, backward prediction mode, or bidirectional prediction mode) and the motion compensation error correction information in the video memory 54 based on the motion vector information. The data subjected to the motion compensation becomes an error correction value in the spatial domain. This error correction value is supplied to the discrete cosine transform device 56.

The discrete cosine transform unit 56 performs a discrete cosine transform on the supplied error correction value, and generates a motion compensation error correction coefficient which is an error correction value in the frequency domain. The motion compensation error correction coefficient is supplied to the adder 40.

In the adder 40, the error due to motion compensation is corrected by subtracting the motion compensation error correction coefficient from the discrete cosine transform coefficient inversely quantized by the inverse quantizer 14. You.

In the image information conversion device 30 according to the third embodiment to which the present invention is applied, data of each block is transferred in the frequency domain to reduce the code amount (bit rate). Therefore, the amount of calculation can be reduced and the circuit configuration can be significantly reduced as compared with a conventional image information conversion device that decodes and encodes baseband video data. At the same time, the image information conversion device 30 can reduce the code amount without causing image quality deterioration due to accumulation of motion compensation errors.

The inverse discrete cosine transform device 53 and the discrete cosine transform device 5 of the motion compensation error correction device 50 are used.
6, reference "A fast computation"
algorithmic for the discr
ETTE COSINE TRANSFORM ”(IEEE
E Trans. Commun. , Vol. 25, n
o. 9 pp. 1004-1009, 1977).

In the inverse discrete cosine transform device 53 and the discrete cosine transform device 56, when the coefficient of the horizontal high frequency component is replaced with 0 in the band limiting device 15,
By omitting the inverse discrete cosine transform and the discrete cosine transform for the coefficient replaced with 0, it is possible to reduce the circuit scale and the amount of arithmetic processing.

Further, the deterioration of the color difference signal in the image is as follows.
Compared to the degradation of the luminance signal, it has a characteristic that it is difficult for the human eye to understand.By applying the above-described motion compensation error correction only to the luminance signal, the circuit scale and The amount of arithmetic processing can be greatly reduced. Further, the error in the P picture propagates to the B picture, but the error in the B picture does not propagate any more. On the other hand, a B picture includes a bidirectional prediction mode and requires a large amount of calculation processing. Therefore, by performing the motion compensation error correction only on the P picture, it is conceivable to significantly reduce the circuit scale and the amount of arithmetic processing while keeping the image quality deterioration to a minimum. By not performing the processing on the B picture, the capacity of the video memory 54 can be reduced.

As described above, the third embodiment to which the present invention is applied
In the image information conversion device 30 according to the embodiment of the present invention, the amount of code (bit rate) can be reduced by exchanging data of each block in the frequency domain. Compared with the conventional image information conversion apparatus, the amount of calculation is reduced, and the circuit configuration can be significantly reduced.

In the image information conversion device 30 according to the third embodiment to which the present invention is applied, the quantization matrix used in the quantization device 18 is changed from the quantization matrix for the intra macroblock to the intra macroblock. By switching to a quantization matrix for inter-macroblocks that does not coarsely quantize high-frequency components as compared to a quantization matrix for image quality degradation in I-pictures is prevented, and the image flash phenomenon is also subjectively avoided. Accordingly, the image quality of the P picture and the B picture formed based on the I picture can be improved.

Further, in the image information conversion apparatus 30 according to the third embodiment to which the present invention is applied, the quantization matrix for an inter macro block is converted into both an intra macro block and an inter macro block. By using this, the quantization matrix switching device 17 does not need to include a storage medium and store the quantization matrix for switching.

[0192] In the above-described image information conversion apparatus 30, image compression information (bit stream) according to MPEG-2 is input, but image compression information (bit stream) encoded by orthogonal transformation and motion compensation is used. If there is, for example, MPEG-1 or H.264. 263 or the like may be input.

[0193]

As described above, according to the image information conversion apparatus and the image information conversion method according to the present invention, the quantization matrix used in the encoding means or the quantization means is replaced with the quantization matrix for the intra macroblock. To the quantization matrix for the inter macroblock, the picture quality deterioration in the I picture is prevented, and the flash phenomenon of the image is also subjectively avoided. Also, the image quality of the B picture can be improved.

Further, according to the image information conversion apparatus and the image information conversion method according to the present invention, the quantization matrix for the inter macro block is used for both the intra macro block and the inter macro block. ,
There is no need to provide a storage medium in the quantization matrix switching means and store the quantization matrix for switching.

[Brief description of the drawings]

FIG. 1 is a block diagram of an image information conversion apparatus according to a first embodiment of the present invention.

FIG. 2 is a diagram showing default values of a quantization matrix.
(A) is a figure which shows the quantization matrix set by default used about an intra macroblock,
(B) is a figure which shows the quantization matrix set to the default value used about an inter macroblock.

FIG. 3 is a diagram illustrating a quantization matrix for encoding an intra macroblock.

FIG. 4 is a diagram illustrating a transition of a signal-to-noise ratio of a luminance signal with respect to an original image of image compression information whose code amount has been reduced by the image information conversion device according to the first embodiment of the present invention; .

FIG. 5 is a diagram illustrating the pSNR of a luminance signal when a quantization matrix is switched and when a quantization matrix is not switched in an image information conversion apparatus according to a first embodiment of the present invention. FIG. 5 is a diagram showing a change in a measurement result represented by “”.

FIG. 6 is a diagram illustrating default values of a quantization matrix.
(A) is a figure which shows the quantization matrix set by default used about an intra macroblock,
(B) is a diagram showing a quantization matrix set to a default value used for an inter macroblock,
(C) is a diagram showing a quantization matrix defined by Test Model 5.

FIG. 7 is a diagram illustrating a quantization matrix for an intra macroblock used for output image compression information.

FIG. 8 is a diagram illustrating a code amount assigned to each frame before and after switching of a quantization matrix in the image information conversion apparatus according to the first embodiment to which the present invention is applied.

FIG. 9 is a block diagram of an image information conversion apparatus according to a second embodiment of the present invention.

FIG. 10 is a diagram showing a scan order of discrete cosine transform coefficients when performing variable length coding. (A) is a figure which shows the scan order of a zigzag scan, (b) is a figure which shows the scan order of an alternate scan.

FIG. 11 is a diagram illustrating an example of band limitation of a horizontal high-frequency component of a discrete cosine transform coefficient by the band limitation device of the image information conversion device according to the second embodiment. (A) is a figure which shows the example of a band limitation of the discrete cosine transform coefficient with respect to a luminance signal, (b) is a figure which shows the example of a band limitation of the discrete cosine transform coefficient with respect to a chrominance signal.

FIG. 12 is a flowchart illustrating the operation of a code amount control device of the image information conversion device according to the second embodiment.

FIG. 13 is a diagram illustrating a configuration of a pseudo GOP.

FIG. 14 is a diagram illustrating scanning of a discrete cosine transform coefficient by an alternate scan method.
(A) is a figure which shows the discrete cosine transform coefficient before band limitation, (b) is a figure which shows the discrete cosine transform coefficient after band limitation.

FIG. 15 is a diagram illustrating a transition of a signal-to-noise ratio of a luminance signal with respect to an original image of image compression information whose code amount has been reduced by the image information conversion apparatus according to the second embodiment of the present invention. .

FIG. 16 illustrates an image information conversion device according to a second embodiment to which the present invention is applied, in a case where the quantization matrix is switched and in a case where the quantization matrix is not switched.
FIG. 9 is a diagram illustrating a change in a measurement result represented by a pSNR of a luminance signal.

FIG. 17 is a diagram illustrating a code amount assigned to each frame before and after switching of a quantization matrix in the image information conversion device according to the second embodiment of the present invention;

FIG. 18 is a block diagram of an image information conversion apparatus according to a third embodiment of the present invention.

FIG. 19 is a block diagram of a conventional image information conversion device.

FIG. 20 is a block diagram of a conventional image information conversion device.

FIG. 21 is a diagram showing a transition of a signal-to-noise ratio of a luminance signal with respect to an original image of image compression information encoded by a conventional image information conversion device.

FIG. 22 is a diagram showing a transition of a signal-to-noise ratio of a luminance signal with respect to an original image in image compression information encoded by a conventional image information conversion device.

FIG. 23 is a diagram illustrating default values of a quantization matrix.
(A) is a figure which shows the quantization matrix set by default used about an intra macroblock,
(B) is a diagram showing a quantization matrix set to a default value used for an inter macroblock,
(C) is a diagram showing a quantization matrix defined by Test Model 5.

FIG. 24 is a diagram illustrating a transition of a signal-to-noise ratio of a luminance signal with respect to an original image of image compression information whose code amount has been reduced by a conventional image information conversion device.

FIG. 25 is a diagram illustrating a transition of a signal-to-noise ratio of a luminance signal with respect to an original image of image compression information whose code amount has been reduced by a conventional image information conversion device.

[Explanation of symbols]

Reference Signs List 1 image information conversion device, 2 image information decoding device, 3 additional information buffer, 4 quantization matrix switching device, 5 image information encoding device, 10 image information conversion device, 11 code buffer, 12 compression information analysis device, 13 variable length Decoding device, 14 inverse quantization device, 15 band limiting device, 16
Information buffer, 17 quantization matrix switching device, 18 quantization device, 19 variable length coding device, 20 code buffer, 21 code amount control device

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) 9A001 (72) Inventor Shintaro Okada 6-7-35 Kita Shinagawa, Shinagawa-ku, Tokyo Inside Sony Corporation (72 ) Inventor Ryu Ik 6-7-35 Kita-Shinagawa, Shinagawa-ku, Tokyo, Japan Sony Corporation (72) Inventor Naofumi Yanagihara 6-35, Kita-Shinagawa, Shinagawa-ku, Tokyo Sony Corporation F-term (reference) ) 5C052 AA17 AB03 AB04 CC11 DD10 GB06 GC05 GE04 5C055 AA07 EA02 EA04 EA16 FA22 HA31 5C059 KK01 KK41 MA00 MA05 MA14 MA23 MC14 ME01 NN21 PP05 PP06 PP07 PP16 RC14 RC16 SS02 SS06 SS12 SS13 TA43 TA47 TB03 TC02 TC06 TC03 UA11 5C066 AA02 AA06 AA07 BA17 CA05 DD01 DD06 EF11 GA02 GA05 HA02 KC08 KE01 KE03 KE09 KF05 5C078 BA21 BA44 BA53 DA01 9A001 EE02 EZ04 HZ27

Claims (10)

[Claims] A first image compression information of a first bit rate obtained by compression-encoding an image signal by performing an orthogonal transformation in units of an orthogonal transformation block including a predetermined pixel block, based on the first bit rate. An image information conversion device for converting the first image compression information into a second image compression information having a lower bit rate and a second bit rate, the decoding means generating the moving image information by decoding the first image compression information; The intra-macro block quantization matrix, which is a quantization matrix for intra-frame coding used when image compression information is generated, is used for an inter-macro block, which is a quantization matrix for inter-frame coding. A quantization matrix switching unit for switching to a quantization matrix; and a decoding unit based on additional information used when the decoding unit decodes the first image compression information. Encoding means for orthogonally transforming the obtained moving image information to encode the second image compression information, wherein the encoding means comprises a quantizer for the inter macro block switched by the quantization matrix switching means. Quantizing the intra macro block based on the quantization matrix, and quantizing the inter macro block based on the quantization matrix for the inter macro block used when the decoding unit decodes the first image compression information. An image information conversion device characterized by the above-mentioned. 2. The quantization matrix switching means switches the intra-macroblock quantization matrix to an inter-macroblock quantization matrix, and selects a (0, 0, 1) of the switched inter-macroblock quantization matrix. 2. The image information conversion device according to claim 1, wherein when the (0) component is not 8, a quantization matrix obtained by converting the (0,0) component into 8 is generated. 3. The first image compression information of a first bit rate obtained by compressing and encoding an image signal by performing an orthogonal transformation in units of an orthogonal transformation block composed of a predetermined pixel block, based on the first bit rate. An image compression apparatus for converting the input first image compression information into an image information conversion device for converting the first image compression information into a second image compression information having a lower bit rate. An inverse quantization means for inversely quantizing the orthogonal transform coefficient of the first image compression information input to the image compression information analysis means, based on analysis result information which is an analysis result analyzed by the analysis means; An orthogonal transformation of the first image compression information inversely quantized by the inverse quantization means based on a quantization width such that the second image compression information becomes the second bit rate. A quantization means for requantizing the permutation coefficient, and a quantization matrix for an intra macroblock, which is a quantization matrix for intra-frame encoding used when the first image compression information is generated, And a quantization matrix switching means for switching to a quantization matrix for an inter macroblock which is a quantization matrix for inter-coding, wherein the quantization means is for an inter macroblock switched by the quantization matrix switching means. Quantizing an intra macroblock based on a quantization matrix and the quantization width, and a quantization matrix for an inter macroblock used when the inverse quantization means inversely quantizes the first image compression information. An image information conversion apparatus, wherein an inter macro block is quantized based on the quantization width. 4. The quantization matrix switching means switches the quantization matrix for the intra macro block to a quantization matrix for an inter macro block, and the (0, 0, 1) of the switched quantization matrix for the inter macro block. 4. The image information conversion device according to claim 3, wherein when the (0) component is not 8, a quantization matrix obtained by converting the (0,0) component into 8 is generated. 5. A band limiting means for limiting a value of a high frequency component in a horizontal direction of the orthogonal transform coefficient inversely quantized by the inverse quantization means, wherein the quantization means is switched by the quantization matrix switching means. 4. The orthogonal transform coefficient whose horizontal high-frequency component is restricted by said band-limiting means, based on said inter-macroblock quantization matrix and said quantization width. Image information conversion device. 6. A first image compression information of a first bit rate obtained by compressing and encoding an image signal by performing an orthogonal transformation in units of an orthogonal transformation block composed of a predetermined pixel block, based on the first bit rate. An image information conversion method for converting the first image compression information into a second image compression information having a lower bit rate, generating the moving image information by decoding the first image compression information, A quantization matrix for intra macroblocks, which is a quantization matrix for intraframe coding used when information is generated, and a quantization matrix for inter macroblocks, which is a quantization matrix for interframe coding. Based on the additional information used when decoding the first image compression information and the switched quantization matrix for the inter macroblock. An image characterized by quantizing a black block and quantizing the inter macro block based on the additional information and a quantization matrix for the inter macro block used when decoding the first image compression information. Information conversion method. 7. The quantization matrix for an intra macro block is switched to a quantization matrix for an inter macro block, and the (0,0) component of the switched quantization matrix for an inter macro block is not 8, 7. The image information conversion method according to claim 6, further comprising: generating a quantization matrix obtained by converting the (0, 0) component into 8. 8. First image compression information of a first bit rate obtained by compressing and encoding an image signal by performing an orthogonal transformation in units of an orthogonal transformation block composed of a predetermined pixel block, based on the first bit rate. The image information conversion method of converting the first image compression information of the first bit rate into the second image compression information of the second bit rate having a lower bit rate. The image compression information is analyzed, and the orthogonal transform coefficient of the input first image compression information is inversely quantized based on the analysis result information that is the analysis result, and the first image compression information is generated. The quantization matrix for intra macroblocks, which is the quantization matrix for intra-frame coding used when encoding, is changed to the quantity for inter-macroblocks, which is the quantization matrix for interframe coding. Based on the switched quantization matrix for the inter macroblock and the quantization width at which the output second image compression information becomes the second bit rate. And quantizing the inter macroblock based on the quantization matrix for the inter macro block used when dequantizing the first image compression information and the quantization width. Image information conversion method. 9. The quantization matrix for the intra macro block is switched to a quantization matrix for an inter macro block, and the (0,0) component of the switched quantization matrix for the inter macro block is not 8, 9. The image information conversion method according to claim 8, wherein said generating a quantization matrix by converting said (0,0) component into 8. 10. The inverse-quantized orthogonal transform coefficient limits the value of a high-frequency component in the horizontal direction, and selects a high-frequency signal in the horizontal direction based on the switched quantization matrix for the inter macroblock and the quantization width. 9. The image information conversion method according to claim 8, wherein the orthogonal transform coefficients having limited components are requantized.