JP4983908B2 - Decoding apparatus and method - Google Patents

Description

The present invention relates to MPEG (Moving Picture Experts Group), H.264. Image compression information (bitstream) compressed by orthogonal transformation such as discrete cosine transformation or Karhunen-Labe transformation and motion prediction / compensation, such as 26x, is transmitted via network media such as satellite broadcasting, cable TV, or the Internet. upon receiving Te, or optical disk, to a decoding apparatus and method of the image information used in processing on a storage medium such as a magnetic disk or a flash memory.

  In recent years, image information has been handled as digital data. At that time, for the purpose of efficient transmission and storage of information, compression is performed by orthogonal transform such as discrete cosine transform and motion prediction / compensation using redundancy unique to image information. An apparatus conforming to a system such as MPEG is becoming widespread in both information distribution in broadcasting stations and information reception in general households.

  In particular, MPEG2 (ISO / IEC 13818-2) is defined as a general-purpose image coding system, and is a standard that covers both interlaced and progressively scanned images, standard resolution images, and high-definition images. And widely used in a wide range of applications for consumer use. By using the MPEG2 compression method, for example, a standard resolution interlaced scanning image having 720 × 480 pixels is 4 to 8 Mbps, and a high resolution interlaced scanning image having 1920 × 1088 pixels is 18 to 22 Mbps. (Bit rate) can be assigned to achieve a high compression rate and good image quality.

  MPEG2 was mainly intended for high-quality encoding suitable for broadcasting, but did not support encoding methods with a lower code amount (bit rate) than MPEG1, that is, a higher compression rate. However, with the widespread use of mobile terminals, the need for such an encoding system is expected to increase in the future, and the MPEG4 encoding system has been standardized accordingly. Regarding the image coding system, the standard was approved as an international standard as ISO / IEC 14496-2 in December 1998.

  Furthermore, in recent years, the standardization of a standard called H.264 (ITU-T Q6 / 16 VCEG) has been advanced with the initial purpose of image coding for video conferencing. H. H.264 is known to achieve higher encoding efficiency than a conventional encoding method such as MPEG2 or MPEG4, although a larger amount of calculation is required for encoding and decoding. In addition, as part of MPEG4 activities, this H.264 H.264, H.264 Standardization that incorporates functions not supported by H.264 and achieves higher coding efficiency is performed by JVT (Joint Video Team).

  Here, FIG. 9 shows a schematic configuration of an image information encoding apparatus that realizes image compression by orthogonal transformation such as discrete cosine transformation or Karhunen-Labe transformation and motion prediction / compensation. As illustrated in FIG. 9, the image information encoding apparatus 100 includes an A / D (Analogue / Digital) conversion unit 101, an image rearrangement buffer 102, an adder 103, an orthogonal conversion unit 104, and a quantization unit 105. A lossless encoding unit 106, an accumulation buffer 107, an inverse quantization unit 108, an inverse orthogonal transform unit 109, an adder 110, a frame memory 111, a motion prediction / compensation unit 112, and an intra prediction unit 113. And the rate control unit 114.

  In FIG. 9, an A / D converter 101 converts an input image signal into a digital signal. Then, the image rearrangement buffer 102 rearranges the frames according to the GOP (Group of Pictures) structure of the compressed image information output from the image information encoding device 100. Here, the image rearrangement buffer 102 supplies the image information of the entire frame to the orthogonal transform unit 104 regarding the image on which intra (intra-image) encoding is performed. The orthogonal transform unit 104 performs orthogonal transform such as discrete cosine transform or Karhunen-Loeve transform on the image information, and supplies the transform coefficient to the quantization unit 105. The quantization unit 105 performs a quantization process on the transform coefficient supplied from the orthogonal transform unit 104.

  The lossless encoding unit 106 performs lossless encoding such as variable length encoding and arithmetic encoding on the quantized transform coefficient, and supplies the encoded transform coefficient to the accumulation buffer 107 for accumulation. The encoded transform coefficient is output as image compression information.

  The behavior of the quantization unit 105 is controlled by the rate control unit 114. The quantization unit 105 supplies the quantized transform coefficient to the inverse quantization unit 108, and the inverse quantization unit 108 inversely quantizes the transform coefficient. The inverse orthogonal transform unit 109 performs inverse orthogonal transform processing on the inversely quantized transform coefficients to generate decoded image information, and supplies the information to the frame memory 111 for accumulation.

  On the other hand, the image rearrangement buffer 102 supplies image information to the motion prediction / compensation unit 112 for an image on which inter (inter-image) encoding is performed. The motion prediction / compensation unit 112 extracts image information that is referred to at the same time from the frame memory 111 and performs motion prediction / compensation processing to generate reference image information. The motion prediction / compensation unit 112 supplies the reference image information to the adder 103, and the adder 103 converts the reference image information into a difference signal from the image information. In addition, the motion compensation / prediction unit 112 supplies motion vector information to the lossless encoding unit 106 at the same time.

  The lossless encoding unit 106 performs lossless encoding processing such as variable length encoding or arithmetic encoding on the motion vector information, and forms information to be inserted into the header portion of the image compression information. The other processing is the same as the image compression information subjected to intra coding, and thus description thereof is omitted.

  Here, in the above-described encoding method standardized by JVT (hereinafter referred to as JVT Codec), when performing intra encoding, a prediction image is generated from pixels around a block and the difference is encoded. Intra-prediction coding is adopted. That is, for an image to be intra-encoded, a predicted image is generated from pixel values that have already been encoded in the vicinity of the pixel block to be encoded, and a difference from the predicted image is encoded. The inverse quantization unit 108 and the inverse orthogonal transform unit 109 respectively perform inverse quantization and inverse orthogonal transform on the intra-coded pixels, and the adder 110 encodes the output of the inverse orthogonal transform unit 109 and the pixel block. The predicted image used at the time is added, and the added value is supplied to the frame memory 111 and accumulated. In the case of a pixel block to be intra-encoded, the intra prediction unit 113 reads out neighboring pixels that have already been encoded and accumulated in the frame memory 111, and generates a predicted image. At this time, the lossless encoding unit 106 also performs lossless encoding processing on the intra prediction mode used for generating the predicted image, and outputs the image by including it in the compressed image information.

  Next, FIG. 10 shows a schematic configuration of an image information decoding apparatus corresponding to the image information encoding apparatus 100 described above. As illustrated in FIG. 10, the image information decoding device 120 includes a storage buffer 121, a lossless decoding unit 122, an inverse quantization unit 123, an inverse orthogonal transform unit 124, an adder 125, and an image rearrangement buffer 126. , A D / A (Digital / Analogue) conversion unit 127, a motion prediction / compensation unit 128, a frame memory 129, and an intra prediction unit 130.

  In FIG. 10, the accumulation buffer 121 temporarily stores input image compression information, and then transfers it to the lossless decoding unit 122. The lossless decoding unit 122 performs processing such as variable length decoding or arithmetic decoding on the compressed image information based on the determined format of the compressed image information, and supplies the quantized transform coefficient to the inverse quantization unit 123. . Further, when the frame is inter-coded, the lossless decoding unit 122 also decodes the motion vector information stored in the header portion of the image compression information, and the information is motion prediction / compensation unit 128. To supply.

  The inverse quantization unit 123 inversely quantizes the quantized transform coefficient supplied from the lossless decoding unit 122 and supplies the transform coefficient to the inverse orthogonal transform unit 124. The inverse orthogonal transform unit 124 performs inverse orthogonal transform such as inverse discrete cosine transform or inverse Karhunen-Loeve transform on the transform coefficient based on the determined format of the image compression information.

  Here, when the frame is intra-coded, the image information subjected to the inverse orthogonal transform process is stored in the image rearrangement buffer 126 and is subjected to D / A conversion in the D / A conversion unit 127. Output after processing.

  On the other hand, when the frame is inter-coded, the motion prediction / compensation unit 128 refers to the motion vector information subjected to the lossless decoding process and the image information stored in the frame memory 129. An image is generated and supplied to the adder 125. The adder 125 synthesizes the reference image and the output of the inverse orthogonal transform unit 124. The other processing is the same as that of the intra-encoded frame, and thus description thereof is omitted.

  Since JVT Codec employs intra-prediction coding, if the frame is intra-coded, the intra-prediction unit 130 reads an image from the frame memory 129, and the lossless decoding unit 122. A prediction image is generated according to the intra prediction mode that has been subjected to the lossless decoding process. The adder 125 adds the output of the inverse orthogonal transform unit 124 and the predicted image.

  The image information encoding device 100 and the image information decoding device 120 described above are described in, for example, Patent Documents 1 and 2 listed below.

JP 2001-199818 A JP 2002-20953 A

  By the way, in JVT Codec (H.264 | MPEG-4 AVC), as described above, when performing intra coding, intra prediction in which a prediction image is generated from pixels around a block and the difference is encoded. Encoding is employed.

  Here, regarding the luminance component, there are two predictions: an intra 4 × 4 prediction mode in which prediction is performed in units of 4 × 4 pixel blocks and an intra 16 × 16 prediction mode in which prediction is performed in units of 16 × 16 pixel blocks (macroblocks). The method is used.

On the other hand, the color difference component is predicted in units of 8 × 8 blocks of Cb and Cr. This predictive encoding method is the same as the intra 16 × 16 prediction mode, and the prediction mode is changed to 8 × 8 block units. FIG. 11 shows a prediction mode in intra prediction coding of color differences. As shown in Figure 11, JVT Codec
(a) Vertical mode (mode = 0)
(b) Horizontal mode (mode = 1)
(c) DC mode (mode = 2)
(d) Plane Prediction mode (mode = 3)
These four prediction modes are defined, and a prediction image is generated according to the prediction mode with the smallest prediction residual. Hereinafter, a method for generating a predicted image in the four prediction modes will be described.

(a) Vertical mode (mode = 0)
In the vertical mode, the pixel of the upper block adjacent to the color difference block (the upper macro block in the case of 4: 2: 0 format) is copied and used as the predicted image of the block. The predicted image pred _c of the color difference block in this case is represented by the following equation (1), where p [x, −1] is the pixel of the adjacent upper block. This mode can be used only when there is an adjacent upper block.

(b) Horizontal mode (mode = 1)
In the horizontal mode, the pixel of the left block adjacent to the chrominance block (left macro block in the case of 4: 2: 0 format) is copied and used as the predicted image of the block. The predicted image pred _c chrominance block in this case, the pixels of adjacent left side block p [-1, y] When is expressed by the following equation (2). This mode can be used only when there is an adjacent left block.

(c) DC mode (mode = 2)
In DC mode, the average value is used as a predicted image using pixels of the upper and left blocks adjacent to the color difference block. However, if there is no adjacent pixel, the value 128 is used as the prediction signal.

That is, when x, y = 0.0.3, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, -1] and left pixel p [-1, y] (where x, y = 0.3). More specifically, when (i) the pixel p [x, -1] and the pixel p [-1, y] are both present, (ii) the pixel p [x, -1] is present and the pixel p [-1 , Y] does not exist, (iii) pixel p [x, -1] does not exist, and pixel p [-1, y] exists, (iv) pixel p [x, -1] and pixel In the case where p [-1, y] does not exist, the four cases are generated according to the following equations (3) to (6), respectively.

Similarly, when x = 4.7.y and y = 0.0.3, the predicted image pred _c [x, y] is the adjacent upper pixel p [x, −1] and left pixel p [−1. , Y] (where x = 4.7, y = 0.0.3). More specifically, (i) when pixel p [x, -1] exists, (ii) when pixel p [x, -1] does not exist and pixel p [-1, y] exists, iii) When the pixel p [x, −1] and the pixel p [−1, y] do not exist, they are generated according to the following equations (7) to (9), respectively.

Similarly, when x = 0.0.3 and y = 4.7.7, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, −1] and left pixel p [−1. , Y] (where x = 0.0.3, y = 4.7.7). More specifically, (i) when pixel p [-1, y] is present, (ii) when pixel p [x, -1] is present, and pixel p [-1, y] is not present, (iii ) When the pixel p [x, −1] and the pixel p [−1, y] do not exist, they are generated according to the following equations (10) to (12), respectively.

Similarly, when x, y = 4.7.7, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, -1] and left pixel p [-1, y] (provided that , X, y = 4.7.7). More specifically, when (i) the pixel p [x, -1] and the pixel p [-1, y] are both present, (ii) the pixel p [x, -1] is present and the pixel p [-1 , Y] does not exist, (iii) pixel p [x, -1] does not exist, and pixel p [-1, y] exists, (iv) pixel p [x, -1] and pixel In the case where p [−1, y] does not exist, the four cases are generated according to the following equations (13) to (16), respectively.

(d) Plane Prediction mode (mode = 3)
In Plane Prediction mode, the predicted image of the block is obtained by approximating the predicted image from the pixel of the left block adjacent to the chrominance block (left macro block in the case of 4: 2: 0 format) and the pixel of the upper block. And In this case, the color difference prediction image pred _c is expressed by the following equation (17), where p [−1, y] is the pixel of the adjacent left block and p [x, −1] is the upper block. expressed. Here, Clip1 in Expression (17) indicates that clipping is performed in the range of 0 to 255.

As described above, after performing intra prediction of color difference components in any of the four prediction modes to generate a predicted image, the adder 103 generates a difference signal between the current pixel block and the predicted image. The orthogonal transform unit 104 applies 4 × 4 integer transform in units of 4 × 4 pixel blocks to the difference signal of 8 × 8 blocks. When the difference signal obtained by subtracting the predicted image from the current pixel block is F ₄ × 4, 4 × 4 integer conversion is expressed as the following Expression (18).

Furthermore, in JVT Codec, after 4 × 4 integer conversion, as shown in FIG. 12, 4 × 4 block (0,0) coefficients (DC coefficients) in 8 × 8 blocks are collected to form 2 × 2 blocks. And 2 × 2 Hadamard transform is applied to this 2 × 2 block. This is because the efficiency of intra prediction used for color difference is not so high, and there is still a correlation between (0, 0) coefficients between adjacent 4 × 4 blocks. In order to further improve the coding efficiency by using this correlation, only (0,0) coefficients of 4 × 4 blocks are collected to form a 2 × 2 block, and 2 × 2 Hadamard transform is applied. Assuming that the 2 × 2 chroma DC block is fdc ₂ × 2, the chroma DC block fdc ′ ₂ × 2 after the 2 × 2 Hadamard transform is expressed by the following equation (19).

After integer conversion, each coefficient is quantized. Assuming that the parameter for obtaining the luminance quantization coefficient is QP _y , the parameter QP _c for obtaining the color difference quantization coefficient is calculated as follows.

That is, first, the parameter QP _i is calculated according to the following equation (20) using the QP _y encoded in the image compression information (takes a value from 0 to 51) and the offset value chroma_qp_offset of the color difference quantization coefficient. . However, QP _i is clipped in the range of 0 to 51.

Then, using this QP _i , a color difference parameter QP _c is obtained from the table shown in Table 1 below.

  Here, assuming that the value of each AC coefficient before quantization is f and the value of each AC coefficient after quantization is f ′, the value of the coefficient after quantization is expressed by the following equation (21).

  On the other hand, assuming that the value of each DC coefficient before quantization is fdc and the value of each DC coefficient after quantization is fdc ′, the value of the coefficient after quantization is expressed by the following equation (22). In the equation (22), r is a constant for rounding processing.

  Further, the inverse quantization of the AC coefficient is represented by the following equation (23), where the AC coefficient after the inverse quantization is f ″.

On the other hand, the inverse quantization of DC coefficients, when the DC coefficient after the inverse quantization and fdc ", if QP _c is 6 or more is represented by the following formula (24), QP _c is less than 6 In some cases, it is expressed by the following equation (25).

  As described above, intra prediction encoding is performed in JVT Codec, but even if the above method is used, intra prediction encoding for color difference is small in block size, so encoding efficiency is not good compared to luminance. There was a problem.

  Further, the above method is compatible only with the 4: 2: 0 format and the YCbCr color space, and in the case of the 4: 2: 2 format, the 4: 4: 4 format, the RGB color space, the XYZ color space, and the like. There was a problem that could not be encoded.

The present invention has been proposed in view of such a conventional situation, and even in an image of 4: 2: 2 format, 4: 4: 4 format, RGB color space, XYZ color space, etc. and to provide a decoding apparatus and method of the image information that enables efficient coding.

  In order to achieve the above-described object, an image information encoding apparatus according to the present invention blocks an input image signal, performs orthogonal transform in units of the block, and performs quantization in the image information encoding apparatus. Intra-picture prediction means for adaptively changing the block size based on a chroma format signal indicating resolution and a color space signal indicating color space, and generating a prediction image when the color signal is subjected to intra-picture prediction encoding, A conversion unit that converts the difference signal between the prediction image generated by the intra-picture prediction unit and the original image into integers in units of a predetermined block size, and adaptively changes the quantization method according to the conversion by the conversion unit, A quantizing means for quantizing the transform coefficient generated by the transform means; the transform coefficient quantized by the quantizing means; the chroma format signal; And a coding means for coding the spatial signal.

  In order to achieve the above-described object, an image information encoding method according to the present invention is an image information encoding method in which an input image signal is blocked and subjected to orthogonal transform in units of the block to perform quantization. An intra-picture prediction step for adaptively changing the block size based on a chroma format signal indicating the resolution of the signal and a color space signal indicating the color space, and generating a prediction image when the color signal is subjected to intra-picture prediction encoding; A conversion step for converting the difference signal between the prediction image generated in the intra-image prediction step and the original image into integers in a predetermined block size unit, and adaptively changing the quantization method according to the conversion in the conversion step A quantization step for quantizing the conversion coefficient generated in the conversion step, the conversion coefficient quantized in the quantization step, the chroma format signal, and the color space. And a coding step for coding a signal.

  In such an image information encoding apparatus and method, when the input image signal is subjected to intra prediction encoding, the resolution of the color component is, for example, 4: 2: 0 format, 4: 2: 2 format, 4: 4: Adapts the block size when generating a predicted image based on a chroma format signal indicating which of the four formats is used and a color space signal indicating which of the color spaces is, for example, YCbCr, RGB, XYZ, etc. Change. Furthermore, in this image information encoding apparatus and method, the chroma format signal and the color space signal are encoded together with the quantized transform coefficient.

In order to achieve the above-described object, the decoding apparatus according to the present invention provides a predicted image generated by intra prediction when a chroma format signal indicating the resolution of a color difference signal of an image signal is in a 4: 4: 4 format. The scale value of the inverse orthogonal transform for the chroma DC block of 4 × 4 blocks obtained by collecting only the DC components of 4 × 4 blocks obtained by performing integer precision orthogonal transform in units of 4 × 4 pixel blocks on the difference signal between the original image and the original image in response calculated by using the quantization scale value when inverse quantization coefficients of chroma DC block quantized, and inverse quantizing means for inverse quantizing the coefficients of chroma DC block quantized, configured to collect only the DC component of the integer precision orthogonal transformed 4 × 4 block in 4 × 4 pixel block inverse quantized transform coefficients by the inverse quantization unit as a target A transform coefficient obtained by orthogonally transforming a 4 × 4 block of chroma DC blocks in units of 4 × 4 pixel blocks is subjected to inverse orthogonal transform in units of 4 × 4 blocks, and the coefficients subjected to inverse orthogonal transform are converted to 8 Inverse transform means for performing inverse orthogonal transform in units of 4 × 4 blocks as each DC coefficient of 16 4 × 4 blocks corresponding to 16 × 16 pixel blocks in which × 8 pixel blocks are arranged in the vertical and horizontal directions, and the above inverse transform using coefficients converted by means, 8 × in 16 × 16 pixel block formed by arranging 8 pixel block in the vertical transverse direction, the image prediction to generate a prediction image when image prediction decoding the chrominance signal Means.

In order to achieve the above-described object, the decoding method according to the present invention is generated by intra-picture prediction when the chroma format signal indicating the resolution of the color difference signal of the image signal is in the 4: 4: 4 format . Scale of inverse orthogonal transform for chroma DC block of 4 × 4 block, which is a collection of only 4 × 4 block DC components obtained by performing integer-precision orthogonal transform in units of 4 × 4 pixel blocks on the difference signal between the predicted image and the original image were calculated according to the value, using the quantization scale value when inverse quantization coefficients of chroma DC block quantized, inverse quantization step for inversely quantizing the coefficients of chroma DC block quantized When configuration gathered only DC components of the 4 × 4 blocks integer precision orthogonal transformation in 4 × 4 pixel block transform coefficients inverse quantized by the inverse quantization step as the target The transform coefficient obtained by orthogonally transforming the 4 × 4 block chroma DC block in units of 4 × 4 pixel blocks is subjected to inverse orthogonal transform in units of 4 × 4 blocks, and the coefficients subjected to inverse orthogonal transform are obtained. , An inverse transform step of performing inverse orthogonal transform in units of 4 × 4 blocks as DC coefficients of 16 4 × 4 blocks corresponding to 16 × 16 pixel blocks in which 8 × 8 pixel blocks are arranged in the vertical and horizontal directions; using the conversion coefficients by the inverse transform step, in 16 × 16 pixel block by arranging the blocks of 8 × 8 pixels in the vertical transverse direction, and generates a prediction image when image prediction decoding the chrominance signal And an intra-picture prediction step.

  In such an image information decoding apparatus and method, a chroma format signal indicating whether the resolution of the color component is, for example, 4: 2: 0 format, 4: 2: 2 format, 4: 4: 4 format, And a color space signal indicating whether the color space is, for example, YCbCr, RGB, XYZ, or the like, and a block signal corresponding to the chroma format signal and the color space signal is used to predictively decode the color signal in the image. A prediction image is generated.

  According to the image information encoding apparatus and method and the image information decoding apparatus and method according to the present invention, the 4: 2: 0 format and YCbCr color space as well as the 4: 2: 2 format and 4: Even in the case of 4: 4 format, RGB color space, XYZ color space, etc., encoding and decoding can be efficiently performed by intra prediction.

It is a figure explaining schematic structure of the picture information coding device in this embodiment. It is a figure which shows an example of a structure of the intra estimation part in the image information encoding device. It is a figure which shows an example of a structure of the orthogonal transformation part in the image information encoding device. It is a figure which shows a mode that the DC coefficient of eight 4x4 blocks in two 8x8 blocks continuous in a vertical direction is collected, and a 2x4 block is comprised. It is a figure which shows an example of a structure of the quantization part in the image information encoding device. It is a figure which shows an example of a structure of the inverse quantization part in the image information encoding device. It is a figure which shows an example of a structure of the inverse orthogonal transformation part in the image information encoding device. It is a figure explaining schematic structure of the image information decoding apparatus in this Embodiment. It is a figure explaining schematic structure of the conventional image information encoding apparatus which implement | achieves image compression by orthogonal transformation, such as discrete cosine transformation or Karhunen-Labe transformation, and motion prediction and compensation. It is a figure explaining schematic structure of the conventional image information decoding apparatus corresponding to the image information encoding apparatus. It is a figure explaining four intra prediction modes in JVT Codec. It is a figure which shows a mode that the DC coefficient of four 4x4 blocks in an 8x8 block is collected, and a 2x2 block is comprised.

  Hereinafter, specific embodiments to which the present invention is applied will be described in detail with reference to the drawings. However, the present invention is not limited to these embodiments, and does not depart from the gist of the present invention. Of course, various changes can be made.

(1) Configuration and Operation of Image Information Encoding Device First, a schematic configuration of an image information encoding device according to the present embodiment is shown in FIG. As illustrated in FIG. 1, the image information encoding device 10 includes an A / D (Analogue / Digital) conversion unit 11, an image rearrangement buffer 12, an adder 13, an orthogonal conversion unit 14, and a quantization unit 15. A lossless encoding unit 16, an accumulation buffer 17, an inverse quantization unit 18, an inverse orthogonal transform unit 19, an adder 20, a frame memory 21, a motion prediction / compensation unit 22, and an intra prediction unit 23. And the rate control unit 24.

  In FIG. 1, an A / D converter 11 converts an input image signal into a digital signal. Then, the image rearrangement buffer 12 performs frame rearrangement according to the GOP (Group of Pictures) structure of the compressed image information output from the image information encoding device 10. Here, the image rearrangement buffer 12 supplies the image information of the entire frame to the orthogonal transform unit 14 regarding the image on which intra (intra-image) encoding is performed. The orthogonal transform unit 14 performs orthogonal transform such as discrete cosine transform or Karhunen-Loeve transform on the image information, and supplies transform coefficients to the quantization unit 15. The quantization unit 15 performs a quantization process on the transform coefficient supplied from the orthogonal transform unit 14.

  The lossless encoding unit 16 performs lossless encoding such as variable length encoding and arithmetic encoding on the quantized transform coefficient, and supplies the encoded transform coefficient to the accumulation buffer 17 for accumulation. The encoded transform coefficient is output as image compression information.

  The behavior of the quantization unit 15 is controlled by the rate control unit 24. In addition, the quantization unit 15 supplies the quantized transform coefficient to the inverse quantization unit 18, and the inverse quantization unit 18 inversely quantizes the transform coefficient. The inverse orthogonal transform unit 19 performs inverse orthogonal transform processing on the inversely quantized transform coefficients to generate decoded image information, and supplies the information to the frame memory 21 for accumulation.

  On the other hand, the image rearrangement buffer 12 supplies image information to the motion prediction / compensation unit 22 for an image on which inter (inter-image) encoding is performed. The motion prediction / compensation unit 22 extracts image information that is referred to at the same time from the frame memory 21 and performs motion prediction / compensation processing to generate reference image information. The motion prediction / compensation unit 22 supplies this reference image information to the adder 13, and the adder 13 converts the reference image information into a difference signal from the image information. In addition, the motion compensation / prediction unit 22 supplies motion vector information to the lossless encoding unit 16 at the same time.

  The lossless encoding unit 16 performs lossless encoding processing such as variable length encoding or arithmetic encoding on the motion vector information, and forms information to be inserted into the header portion of the image compression information. The other processing is the same as the image compression information subjected to intra coding, and thus description thereof is omitted.

  Here, in the above-described JVT Codec, when performing intra coding, intra prediction coding is employed in which a prediction image is generated from pixels around a block and the difference is coded. That is, for an image to be intra-encoded (I picture, I slice, intra macroblock, etc.), a predicted image is generated from pixel values that have already been encoded near the pixel block to be encoded, and the predicted image The difference between is encoded. The inverse quantization unit 18 and the inverse orthogonal transform unit 19 respectively perform inverse quantization and inverse orthogonal transform on the intra-coded pixels, and the adder 20 encodes the output of the inverse orthogonal transform unit 19 and the pixel block. The predicted image used at the time is added, and the added value is supplied to the frame memory 21 and accumulated. In the case of a pixel block to be intra-encoded, the intra prediction unit 23 reads out neighboring pixels that have already been encoded and accumulated in the frame memory 21, and generates a predicted image. At this time, the lossless encoding unit 16 also performs lossless encoding processing on the intra prediction mode used for generating the predicted image, and outputs the image by including it in the compressed image information.

(2) Application Part of the Present Invention in Image Information Encoding Device (2-1) Intra Prediction Unit An example of the configuration of the intra prediction unit 23 is shown in FIG. The intra prediction unit 23 includes a chroma format signal indicating whether the resolution of the color component is 4: 2: 0 format, 4: 2: 2 format, 4: 4: 4 format, and the color space is YCbCr, RGB , XYZ, etc., based on the color space signal indicating the switching, the prediction method is switched. The chroma format signal and the color space signal are set in advance by an external user or the like and supplied to the image information encoding apparatus 10.

  In the intra prediction unit 23 illustrated in FIG. 2, the chroma format signal and the color space signal are supplied to the switches 30 and 32. The switches 30 and 32 select one of the intra predictors 31a, 31b, and 31c based on the chroma format signal and the color space signal, and supply the selected image signal read from the frame memory 21 to the selected intra predictor. A prediction image from the intra predictor that has been output is output. The switches 30 and 32 select the same intra predictor. In FIG. 2, description will be made assuming that one of the three types of intra-predictors 31a, 31b, and 31c is selected. However, the number of intra-predictors, that is, the number of prediction methods, is arbitrarily set. Can do.

(2-1-1)
First, the operation of the intra predictor 31a will be described. In the intra predictor 31a, prediction is performed in units of 8 × 8 blocks on an image signal in which the chroma format signal indicates 4: 2: 0 format and the color space signal indicates YCbCr. The operation of the intra predictor 31a is the same as that of the above-described conventional example, and thus detailed description thereof is omitted.

(2-1-2)
Next, the operation of the intra predictor 31b will be described. In the intra predictor 31b, the intra color difference prediction mode includes four prediction modes of Vertical mode, Horizontal mode, DC mode, and Plane prediction mode. In this intra predictor 31b, two continuous 8 × 8 blocks in a macro block are grouped together for an image signal in which the chroma format signal indicates 4: 2: 2 format and the color space signal indicates YCbCr. An 8 × 16 block is configured, and prediction is performed using the 8 × 16 block as a unit. Hereinafter, a prediction image generation method according to each of the four prediction modes in the intra predictor 31b will be described.

(a) Vertical mode (mode = 0)
In the vertical mode, the pixel of the upper block adjacent to the color difference block is copied and used as the predicted image of the block. In this case, the color difference prediction image pred _c is represented by the following expression (26), where p [x, −1] is the pixel of the adjacent upper block. This mode can be used only when there is an adjacent upper block.

(b) Horizontal mode (mode = 1)
In the horizontal mode, the pixel of the left block adjacent to the color difference block is copied and used as the predicted image of the block. The predicted image pred _c of the color difference block in this case is expressed as the following Expression (27), where p [−1, y] is the pixel of the adjacent left block. This mode can be used only when there is an adjacent left block.

(c) DC mode (mode = 2)
In DC mode, the average value is used as a predicted image using pixels of the upper and left blocks adjacent to the color difference block. However, if there is no adjacent pixel, the value 128 is used as the prediction signal.

That is, when x, y = 0.0.3, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, -1] and left pixel p [-1, y] (where x, y = 0.3). More specifically, when (i) the pixel p [x, -1] and the pixel p [-1, y] are both present, (ii) the pixel p [x, -1] is present and the pixel p [-1 , Y] does not exist, (iii) pixel p [x, -1] does not exist, and pixel p [-1, y] exists, (iv) pixel p [x, -1] and pixel In the case where p [−1, y] does not exist together, they are generated according to the following equations (28) to (31), respectively.

Similarly, when x = 4.7.y and y = 0.0.3, the predicted image pred _c [x, y] is the adjacent upper pixel p [x, −1] and left pixel p [−1. , Y] (where x = 4.7, y = 0.0.3). More specifically, (i) when pixel p [x, -1] exists, (ii) when pixel p [x, -1] does not exist and pixel p [-1, y] exists, iii) When the pixel p [x, −1] and the pixel p [−1, y] do not exist, they are generated according to the following equations (32) to (34), respectively.

Similarly, when x = 0.0.3 and y = 4.7.7, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, −1] and left pixel p [−1. , Y] (where x = 0.0.3, y = 4.7.7). More specifically, (i) when pixel p [-1, y] is present, (ii) when pixel p [x, -1] is present, and pixel p [-1, y] is not present, (iii ) When the pixel p [x, -1] and the pixel p [-1, y] are not present, the pixel p [x, -1] is generated according to the following equations (35) to (37) in the three cases.

Similarly, when x, y = 4.7.7, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, -1] and left pixel p [-1, y] (provided that , X, y = 4.7.7). More specifically, when (i) the pixel p [x, -1] and the pixel p [-1, y] are both present, (ii) the pixel p [x, -1] is present and the pixel p [-1 , Y] does not exist, (iii) pixel p [x, -1] does not exist, and pixel p [-1, y] exists, (iv) pixel p [x, -1] and pixel When p [−1, y] does not exist, the four cases are generated according to the following equations (38) to (41), respectively.

Similarly, when x = 0.0.3 and y = 8.11, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, −1] and left pixel p [−1. , Y] (where x = 0.0.3, y = 8.11). More specifically, (i) when pixel p [-1, y] is present, (ii) when pixel p [x, -1] is present, and pixel p [-1, y] is not present, (iii ) When the pixel p [x, -1] and the pixel p [-1, y] are not present, the pixel p [x, -1] is generated according to the following equations (42) to (44) in the three cases.

Similarly, when x = 4.7.y and y = 8.1.11, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, −1] and left pixel p [−1. , Y] (where x = 4.7.7, y = 8..11). More specifically, when (i) the pixel p [x, -1] and the pixel p [-1, y] are both present, (ii) the pixel p [x, -1] is present and the pixel p [-1 , Y] does not exist, (iii) pixel p [x, -1] does not exist, and pixel p [-1, y] exists, (iv) pixel p [x, -1] and pixel In the case where p [−1, y] does not exist together, they are generated according to the following equations (45) to (48), respectively.

Similarly, when x = 0.0.3 and y = 12.0.15, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, −1] and left pixel p [−1. , Y] (where x = 0.0.3, y = 12.0.15). More specifically, (i) when pixel p [-1, y] is present, (ii) when pixel p [x, -1] is present, and pixel p [-1, y] is not present, (iii ) When the pixel p [x, −1] and the pixel p [−1, y] do not exist, they are generated according to the following equations (49) to (51), respectively.

Similarly, when x = 4.7.y and y = 12.0.15, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, −1] and left pixel p [−1. , Y] (where x = 4.7.7, y = 12.0.15). More specifically, when (i) the pixel p [x, -1] and the pixel p [-1, y] are both present, (ii) the pixel p [x, -1] is present and the pixel p [-1 , Y] does not exist, (iii) pixel p [x, -1] does not exist, and pixel p [-1, y] exists, (iv) pixel p [x, -1] and pixel In the case where p [−1, y] does not exist together, they are generated according to the following equations (52) to (55), respectively.

Here, in the prediction method described above, since the average value of the 8 pixels of the upper block and the 16 pixels of the left block is simply used as the predicted image, division by 24 is necessary and the amount of calculation increases. There is a problem. Therefore, the amount of calculation can be reduced by modifying the prediction method as follows and performing division by 16 (= 2 ⁴ ).

That is, when x, y = 0.0.7, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, -1] and left pixel p [-1, y] (where x, y = 0.7). More specifically, when (i) the pixel p [x, -1] and the pixel p [-1, y] are both present, (ii) the pixel p [x, -1] is present and the pixel p [-1 , Y] does not exist, (iii) pixel p [x, -1] does not exist, and pixel p [-1, y] exists, (iv) pixel p [x, -1] and pixel In the case where p [−1, y] does not exist, the four cases are generated according to the following equations (56) to (59), respectively.

Similarly, when x = 0.0.7 and y = 8..15, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, −1] and left pixel p [−1. , Y] (where x = 0.0.7, y = 8..15). More specifically, (i) when pixel p [-1, y] is present, (ii) when pixel p [x, -1] is present, and pixel p [-1, y] is not present, (iii ) When the pixel p [x, −1] and the pixel p [−1, y] do not exist, they are generated according to the following equations (60) to (62), respectively.

(d) Plane Prediction mode (mode = 3)
In the Plane Prediction mode, a predicted image of the block is approximated by plane approximation from the pixel of the left block adjacent to the color difference block and the pixel of the upper block. In this case, the color difference prediction image pred _c is expressed by the following equation (63), where p [−1, y] is the pixel of the adjacent left block and p [x, −1] is the upper block. expressed. Here, Clip1 in Expression (63) indicates that clipping is performed in the range of 0 to 255.

(2-1-3)
Next, the operation of the intra predictor 31c will be described. Also in the intra predictor 31c, four prediction modes of Vertical mode, Horizontal mode, DC mode, and Plane prediction mode exist in the intra color difference prediction mode. In the intra predictor 31c, four 8 × 8 pixels in the vertical and horizontal directions in the macroblock are continuous with respect to the image signal in which the chroma format signal indicates the 4: 4: 4 format and the color space signal indicates YCbCr, RGB, or XYZ. The blocks are grouped to form a 16 × 16 block, and prediction is performed in units of the 16 × 16 block. Hereinafter, a prediction image generation method according to each of the four prediction modes in the intra predictor 31c will be described.

(a) Vertical mode (mode = 0)
In the vertical mode, the pixel of the upper block adjacent to the color difference block is copied and used as the predicted image of the block. The color difference prediction image pred _{c in} this case is expressed as the following equation (64), where p [x, −1] is the pixel of the adjacent upper block. This mode can be used only when there is an adjacent upper block.

(b) Horizontal mode (mode = 1)
In the horizontal mode, the pixel of the left block adjacent to the color difference block is copied and used as the predicted image of the block. The predicted image pred _c of the color difference block in this case is expressed as the following Expression (65), where p [−1, y] is the pixel of the adjacent left block. This mode can be used only when there is an adjacent left block.

(c) DC mode (mode = 2)
In DC mode, the average value is used as a predicted image using pixels of the upper and left blocks adjacent to the color difference block. However, if there is no adjacent pixel, the value 128 is used as the prediction signal.

That is, when x, y = 0..15, the predicted image pred _c [x, y] includes the adjacent upper pixel p [x, −1] and left pixel p [−1, y] (provided that x, y = 0..15). More specifically, when (i) the pixel p [x, -1] and the pixel p [-1, y] are both present, (ii) the pixel p [x, -1] is present and the pixel p [-1 , Y] does not exist, (iii) pixel p [x, -1] does not exist, and pixel p [-1, y] exists, (iv) pixel p [x, -1] and pixel In the case where p [−1, y] does not exist, the four cases are generated according to the following equations (66) to (69), respectively.

(d) Plane Prediction mode (mode = 3)
In the Plane Prediction mode, a predicted image of the block is approximated by plane approximation from the pixel of the left block adjacent to the color difference block and the pixel of the upper block. Predicted image pred _c chrominance in this case, the pixels of adjacent left side block p [-1, y], the upper block p [x, -1] When, as shown in the following expression (70) expressed. Here, Clip1 in Expression (70) indicates that clipping is performed in the range of 0 to 255.

(2-2) Orthogonal Transform Unit The chroma format signal and the color space signal are also supplied to the orthogonal transform unit 14.
An example of the configuration of the orthogonal transform unit 14 is shown in FIG. The orthogonal transform unit 14 includes a chroma format signal indicating whether the color component resolution is 4: 2: 0 format, 4: 2: 2 format, 4: 4: 4 format, and the color space is YCbCr, RGB. , XYZ, or the like, the orthogonal transform method is switched based on a color space signal indicating whether the signal is XYZ or the like.

  In the orthogonal transform unit 14 shown in FIG. 3, the chroma format signal and the color space signal are supplied to the switches 40 and 42. The switches 40 and 42 select one of the orthogonal transformers 41a, 41b, and 41c based on the chroma format signal and the color space signal, and supply the output from the adder 13 to the selected orthogonal transformer to select the selected orthogonal transformer. The signal from the converter is output. The switches 40 and 42 select the same orthogonal transformer. In FIG. 3, the description will be made assuming that one of the three types of orthogonal transformers 41a, 41b, and 41c is selected. However, the number of the orthogonal transformers, that is, the number of orthogonal transformation schemes is arbitrarily set. be able to.

(2-2-1)
First, the operation of the orthogonal transformer 41a will be described. In the orthogonal transformer 41a, an orthogonal transformation is performed on an image signal in which the chroma format signal indicates 4: 2: 0 format and the color space signal indicates YCbCr. Since the operation of the orthogonal transformer 41a is the same as that of the conventional example described above, detailed description thereof is omitted.

(2-2-2)
Next, the operation of the orthogonal transformer 41b will be described. In the orthogonal transformer 41b, an orthogonal transformation is performed on an image signal in which the chroma format signal indicates 4: 2: 2 format and the color space signal indicates YCbCr.

More specifically, after performing intra prediction of color difference, 4 × 4 integer transform is applied in units of 4 × 4 pixel blocks in 8 × 8 blocks. If the difference signal obtained by subtracting the predicted image from the pixel block is f ₄ × 4, 4 × 4 orthogonal transformation is expressed as the following Expression (71).

After the 4 × 4 integer conversion, as shown in FIG. 4, the 2 × 4 block is configured by collecting the (0,0) coefficients of eight 4 × 4 blocks in two 8 × 8 blocks that are continuous in the vertical direction. The 2 × 4 transform is applied to the 2 × 4 block. This is because the efficiency of intra prediction used for color difference is not so high, and there is still a correlation between (0, 0) coefficients between adjacent 4 × 4 blocks. In order to further increase the coding efficiency by using this correlation, only the (0,0) coefficients of the 4 × 4 block are collected to form a 2 × 4 block, and the 2 × 4 transform is applied. _Assuming that a 2 × 4 chroma DC block is f 2 × 4, the transformation for this chroma DC block is expressed by the following equation (72).

(2-2-3)
Next, the operation of the orthogonal transformer 41c will be described. In the orthogonal transformer 41c, orthogonal transformation is performed on an image signal in which the chroma format signal indicates 4: 4: 4 format and the color space signal indicates YCbCr, RGB, or XYZ.

  More specifically, after the color difference indicating 4: 4: 4 format, YCbCr, RGB, XYZ is converted to 4 × 4 integers, 16 (0, 0) coefficients in the macroblock are collected to obtain 4 in the same manner as the luminance. A x4 DC block is constructed and a 4 x 4 transform is applied. This conversion is expressed as the following Expression (73).

(2-3) Quantization Unit The chroma format signal and the color space signal are also supplied to the quantization unit 15.
An example of the configuration of the quantization unit 15 is shown in FIG. The quantizing unit 15 includes a chroma format signal indicating whether the resolution of the color component is 4: 2: 0 format, 4: 2: 2 format, 4: 4: 4 format, and the color space is YCbCr, RGB. , XYZ, and the like, the quantization method is switched based on a color space signal indicating whether the signal is XYZ or the like.

  In the quantization unit 15 shown in FIG. 5, the chroma format signal and the color space signal are supplied to the switches 50 and 52. In the switches 50 and 52, one of the quantizers 51a, 51b, and 51c is selected based on the chroma format signal and the color space signal, and the output from the orthogonal transform unit 14 is supplied to the selected quantizer. Outputs the signal from the quantizer. The switches 50 and 52 select the same quantizer. In FIG. 5, description is made assuming that one of the three types of quantizers 51a, 51b, and 51c is selected. However, the number of quantizers, that is, the number of quantization methods is arbitrarily set. be able to.

(2-3-1)
First, the operation of the quantizer 51a will be described. The quantizer 51a quantizes an image signal in which the chroma format signal indicates 4: 2: 0 format and the color space signal indicates YCbCr. Since the operation of the quantizer 51a is the same as that of the conventional example described above, detailed description thereof is omitted.

(2-3-2)
Next, the operation of the quantizer 51b will be described. In this quantizer 51b, the chroma format signal indicates the 4: 2: 2 format and the color space signal quantizes the image signal indicating YCbCr.

  Here, the Hadamard transform used for the chroma DC conversion in the case of the 4: 2: 0 format is expressed as the following Expression (74).

  On the other hand, the 2 × 4 conversion used for the conversion of chroma DC in the case of the 4: 2: 2 format is expressed as the following Expression (75).

  Therefore, the normalization coefficient by conversion in the 4: 2: 0 format is ½, whereas the normalization coefficient by conversion in the 4: 2: 2 format is ½√2. However, in this case, since a real number calculation is entered, the calculation is simplified as shown by the following equation (76).

  Since this normalization coefficient is calculated together with the scale at the time of quantization, in the case of conversion in 4: 2: 2 format, it is necessary to change the quantization method as follows.

  If the quantized DC coefficient is Qf ′ [ij], the quantized coefficient value of the 2 × 4 chroma DC block is given by the following equation (77), for example. Here, r in the equation (77) is a parameter for changing the rounding process. Note that the quantization for the AC coefficient is the same as that in the 4: 2: 0 format, and a description thereof will be omitted.

(2-3-3)
Next, the operation of the quantizer 51c will be described. In the quantizer 51c, the chroma format signal indicates the 4: 4: 4 format and the color space signal quantizes the image signal indicating YCbCr, RGB, or XYZ.

  Here, the Hadamard transform used for the chroma DC transform is represented by the following equation (78). Therefore, in this case, the conversion normalization coefficient is 1/4.

  When the quantized DC coefficient is Qf ′ [ij], the quantized coefficient value of the 4 × 4 chroma DC block is given by, for example, the following formula (79). Here, r in Equation (79) is a parameter for changing the rounding process.

(2-4) Inverse Quantization Unit The chroma format signal and the color space signal are also supplied to the inverse quantization unit 18.
An example of the configuration of the inverse quantization unit 18 is shown in FIG. The inverse quantization unit 18 includes a chroma format signal indicating whether the resolution of the color component is 4: 2: 0 format, 4: 2: 2 format, 4: 4: 4 format, and the color space is YCbCr, The inverse quantization method is switched based on a color space signal indicating RGB, XYZ, or the like.

  In the inverse quantization unit 18 shown in FIG. 6, the chroma format signal and the color space signal are supplied to the switches 60 and 62. The switches 60 and 62 select one of the inverse quantizers 61a, 61b, and 61c based on the chroma format signal and the color space signal, and supply the output from the quantization unit 15 to the selected inverse quantizer. Outputs the signal from the selected inverse quantizer. The switches 60 and 62 select the same inverse quantizer. In FIG. 6, the description will be made assuming that one of the three types of inverse quantizers 61a, 61b, and 61c is selected. However, the number of inverse quantizers, that is, the number of inverse quantization methods is arbitrary. Can be set to

(2-4-1)
First, the operation of the inverse quantizer 61a will be described. The inverse quantizer 61a performs inverse quantization on an image signal in which the chroma format signal indicates 4: 2: 0 format and the color space signal indicates YCbCr. Since the operation of the inverse quantizer 61a is the same as that of the conventional example described above, detailed description thereof is omitted.

(2-4-2)
Next, the operation of the inverse quantizer 61b will be described. The inverse quantizer 61b performs inverse quantization on the image signal in which the chroma format signal indicates the 4: 2: 2 format and the color space signal indicates YCbCr.

More specifically, assuming that the DC coefficient after inverse quantization is fdc ″, the coefficient value after inverse quantization of the 2 × 2 chroma DC block is expressed by the following equation (80) when QP _c is 6 or more. When QP _c is less than 6, it is represented by the following formula (81): Since the inverse quantization for the AC coefficient is the same as that in the 4: 2: 0 format, the explanation will be given. Omitted.

(2-4-3)
Next, the operation of the inverse quantizer 61c will be described. The inverse quantizer 61c performs inverse quantization on the image signal in which the chroma format signal indicates 4: 4: 4 format and the color space signal indicates YCbCr, RGB, or XYZ.

More specifically, assuming that the DC coefficient after inverse quantization is fdc ″, the coefficient value after inverse quantization of the 4 × 4 chroma DC block is expressed by the following formula (82) when QP _c is 6 or more. When QP _c is less than 6, it is expressed by the following equation (83): Since the inverse quantization for the AC coefficient is the same as that in the 4: 2: 0 format, the explanation will be given. Omitted.

(2-5) Inverse Orthogonal Transform Unit The chroma format signal and the color space signal are also supplied to the inverse orthogonal transform unit 19.
An example of the configuration of the inverse orthogonal transform unit 19 is shown in FIG. The inverse orthogonal transform unit 19 includes a chroma format signal indicating whether the resolution of the color component is 4: 2: 0 format, 4: 2: 2 format, 4: 4: 4 format, and the color space is YCbCr, The inverse orthogonal transform method is switched based on a color space signal indicating RGB, XYZ, or the like.

  In the inverse orthogonal transform unit 19 shown in FIG. 7, the chroma format signal and the color space signal are supplied to the switches 70 and 72. The switches 70 and 72 select one of the inverse orthogonal transformers 71a, 71b, and 71c based on the chroma format signal and the color space signal, and supply the output from the inverse quantization unit 18 to the selected inverse orthogonal transformer. The signal from the selected inverse orthogonal transformer is output. The switches 70 and 72 select the same inverse orthogonal transformer. In FIG. 7, description is made assuming that one of the three types of inverse orthogonal transformers 71a, 71b, 71c is selected. However, the number of inverse orthogonal transformers, that is, the number of inverse orthogonal transformation methods is arbitrary. Can be set to

(2-5-1)
First, the operation of the inverse orthogonal transformer 71a will be described. In the inverse orthogonal transformer 71a, the inverse orthogonal transform is performed on the image signal in which the chroma format signal indicates the 4: 2: 0 format and the color space signal indicates YCbCr. The operation of the inverse orthogonal transformer 71a is the same as that of the above-described conventional example, and thus detailed description thereof is omitted.

(2-5-2)
Next, the operation of the inverse orthogonal transformer 71b will be described. In the inverse orthogonal transformer 71b, an inverse orthogonal transform is performed on an image signal in which the chroma format signal indicates 4: 2: 2 format and the color space signal indicates YCbCr.

More specifically, the 2 × 4 inverse transform is applied to the 2 × 4 DC block. Assuming that the 2 × 4 chroma DC block after the inverse transform is fdc _{2 × 4} ′ ″, the inverse transform for this chroma DC block is expressed by the following equation (84).

The chroma DC coefficient is set to a (0, 0) coefficient of 4 × 4 blocks as shown in FIG. 4, and inverse conversion of each 4 × 4 block is performed. Each coefficient of a 4 × 4 block having an inversely converted chroma DC fdc _{2 × 4} ′ ″ as a (0,0) coefficient is set to F ′ ₄ × 4, and the decoded differential signal is converted to F ′ 4 × 4. ′ If _{4 × 4} , the inverse transformation is expressed as the following equation (85).

(2-5-3)
Next, the operation of the inverse orthogonal transformer 71c will be described. The inverse orthogonal transformer 71c performs inverse orthogonal transform on an image signal in which the chroma format signal indicates 4: 4: 4 format and the color space signal indicates YCbCr, RGB, or XYZ.

More specifically, the 4 × 4 inverse transform is applied to the 4 × 4 DC block. Assuming that the 4 × 4 chroma DC block after the inverse transform is fdc _{4 × 4} ′ ″, the inverse transform for this chroma DC block is expressed by the following equation (86).

This chroma DC coefficient is set to a (0, 0) coefficient of 4 × 4 blocks of AC coefficients, and inverse conversion of each 4 × 4 block is performed. Each coefficient of the 4 × 4 block having the fdc _{4 × 4} ′ ″ that is the inversely converted chroma DC as the (0,0) coefficient is set to F ′ ₄ × 4, and the decoded differential signal is converted to F ′ 4 × 4. ′ If _{4 × 4} , the inverse transformation is expressed as the following equation (87).

(2-6) Other Blocks The chroma format signal and the color space signal are also supplied to the lossless encoding unit 16, subjected to variable length encoding or arithmetic encoding, are included in the image compression information, and are output.

The chroma format signal and the color space signal are encoded with the following syntax, for example.
seq_parameter_set_rbsp () {
:
chroma_format_idc u (2)
color_space_idc u (2)
:
}
Here, the syntax encoded as u (2) is encoded with a variable-length code such as “001x ₁ x ₀ ”, for example. Of these, x ₁ and x ₀ correspond to 2 bits of syntax to be encoded.

(3) Configuration and Operation of Image Information Decoding Device FIG. 8 shows a schematic configuration of an image information decoding device corresponding to the image information encoding device 10 described above. As shown in FIG. 8, the image information decoding device 80 includes an accumulation buffer 81, a lossless decoding unit 82, an inverse quantization unit 83, an inverse orthogonal transform unit 84, an adder 85, and an image rearrangement buffer 86. , A D / A (Digital / Analogue) conversion unit 87, a motion prediction / compensation unit 88, a frame memory 89, and an intra prediction unit 90.

  In FIG. 8, input image compression information is first stored in the accumulation buffer 81 and then transferred to the lossless decoding unit 82. The lossless decoding unit 82 performs processing such as variable length decoding or arithmetic decoding based on the determined format of the compressed image information. Further, when the frame is inter-coded, the lossless decoding unit 82 also decodes the motion vector information stored in the header portion of the image compression information, and sends the information to the motion prediction / compensation unit 88. Forward. Further, the lossless decoding unit 82 decodes the chroma format signal and the color space signal, and supplies them to the inverse quantization unit 83, the inverse orthogonal transform unit 84, and the intra prediction unit 90.

  The quantized transform coefficient that is output from the lossless decoding unit 82 is supplied to the inverse quantization unit 83, where it is output as a transform coefficient. The inverse orthogonal transform unit 84 performs reversible transform such as inverse discrete cosine transform or inverse Karhunen-Loeve transform based on the format of the determined image compression information. When the frame is intra-coded, the image information subjected to the inverse orthogonal transform process is stored in the image rearrangement buffer 86 and output after the D / A conversion process.

  Here, when the frame or macroblock is intra-coded, the same inverse quantization method and inverse orthogonal transform as described above are performed based on the chroma format signal and the color space signal decoded by the lossless decoding unit 82. And decoding using the intra prediction method.

  On the other hand, if the frame is inter-coded, a reference image is generated based on the motion vector information subjected to the lossless decoding process and the image information stored in the frame memory 89, and the reference image And the output of the inverse orthogonal transform unit 84 are combined in the adder 85. Since other processes are the same as those of the intra-coded frame, description thereof is omitted.

  Intra-picture prediction not only when the input image signal is in 4: 2: 0 format and YCbCr color space but also in 4: 2: 2 format, 4: 4: 4 format, RGB color space, XYZ color space, etc. Encoding can be performed efficiently using encoding.

  DESCRIPTION OF SYMBOLS 10 Image information encoding apparatus, 11 A / D conversion part, 12 Image rearrangement buffer, 13 Adder, 14 Orthogonal transformation part, 15 Quantization part, 16 Lossless encoding part, 17 Storage buffer, 18 Inverse quantization part, 19 inverse orthogonal transform unit, 20 adder, 21 frame memory, 22 motion prediction / compensation unit, 23 intra prediction unit, 24 rate control unit, 80 image information decoding device, 81 storage buffer, 82 lossless decoding unit, 83 inverse quantization , 84 inverse orthogonal transform unit, 85 adder, 86 image rearrangement buffer, 87 D / A conversion unit, 88 motion prediction / compensation unit, 89 frame memory, 90 intra prediction unit

Claims (4)

When the chroma format signal indicating the resolution of the color difference signal of the image signal is in the 4: 4: 4 format, the difference signal between the predicted image generated by intra prediction and the original image is in units of 4 × 4 pixel blocks. The coefficient of the quantized chroma DC block calculated according to the scale value of the inverse orthogonal transform with respect to the 4 × 4 block chroma DC block in which only the DC components of the 4 × 4 block subjected to integer precision orthogonal transform are collected is inverted. using the quantization scale value at the time of quantization, the inverse quantization means for inversely quantizing the coefficients of chroma DC block quantized,
Chroma DC of 4 × 4 block configured to collect only the DC component of the integer precision orthogonal transformed 4 × 4 block in 4 × 4 pixel block inverse quantized transform coefficients by the inverse quantization unit as a target The transform coefficient obtained by performing orthogonal transform in units of 4 × 4 pixel blocks is inversely orthogonal transformed in units of 4 × 4 blocks, and the inversely orthogonally transformed coefficients are converted into 8 × 8 pixel blocks vertically and horizontally. Inverse transform means for performing inverse orthogonal transform in units of 4 × 4 blocks as DC coefficients of 16 4 × 4 blocks corresponding to 16 × 16 pixel blocks arranged in a direction;
Using coefficients converted by the inverse conversion means, in 16 × 16 pixel block by arranging the blocks of 8 × 8 pixels in the vertical transverse direction, and generates a prediction image when image prediction decoding the chrominance signal A decoding device comprising: intra-picture prediction means.   The inverse transform means is a 4 × 4 block unit for a 4 × 4 block chroma DC block configured by collecting only DC components of a 4 × 4 block subjected to integer precision discrete cosine transform in 4 × 4 pixel block units. 16 × 16 obtained by performing inverse discrete Hadamard transform on a 4 × 4 block basis and transforming the inverse discrete Hadamard transform into 8 × 8 pixel blocks in the vertical and horizontal directions. The decoding apparatus according to claim 1, wherein inverse DCT is transformed in units of 4x4 blocks as DC coefficients of transform coefficient blocks corresponding to 16 4x4 pixel blocks constituting a pixel block. When each coefficient of the 4 × 4 chroma DC block is fdc _{4 × 4} , each coefficient fdc ′ _4x4 subjected to the discrete Hadamard transform is

The decoding device according to claim 2 represented by: When the chroma format signal indicating the resolution of the color difference signal of the image signal is in the 4: 4: 4 format, the difference signal between the predicted image generated by intra prediction and the original image is in units of 4 × 4 pixel blocks. The coefficient of the quantized chroma DC block calculated according to the scale value of the inverse orthogonal transform with respect to the 4 × 4 block chroma DC block in which only the DC components of the 4 × 4 block subjected to integer precision orthogonal transform are collected is inverted. An inverse quantization process for inversely quantizing a quantized chroma DC block coefficient using a quantization scale value at the time of quantization ;
Chroma DC of 4 × 4 block configured to collect only the DC component of the integer precision orthogonal transformed 4 × 4 block in 4 × 4 pixel block inverse quantized transform coefficients by the inverse quantization step as the target The transform coefficient obtained by performing orthogonal transform in units of 4 × 4 pixel blocks is inversely orthogonal transformed in units of 4 × 4 blocks, and the inversely orthogonally transformed coefficients are converted into 8 × 8 pixel blocks vertically and horizontally. An inverse transform step of inverse orthogonal transform in units of 4 × 4 blocks as DC coefficients of 16 4 × 4 blocks corresponding to 16 × 16 pixel blocks arranged in a direction;
Using the conversion coefficient by the inverse transform step, in 16 × 16 pixel block by arranging the blocks of 8 × 8 pixels in the vertical transverse direction, generates a prediction image when image prediction decoding the chrominance signal A decoding method comprising: an intra-picture prediction step.

Images