JP2006042187A - Image encoding method, device and program thereof, and image decoding method, device and program thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image conversion method which enables an speed enhancement of a conversion from HD data to SD data. <P>SOLUTION: A form of space area data after an IDCT in image decoding and space area data in image encoding is made a continuation of samples as for a brightness signal. A sample of a first color-difference signal and a sample of a second color-difference signal are made continuous alternatively as for a color-difference signal. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an image encoding method for encoding image data, an image encoding device, an image encoding program, an image decoding method for decoding image data based on the encoded image data, and an image decoding device. The present invention relates to an image decoding program, an image conversion method for converting image data, an image conversion apparatus, and an image conversion program.

  The MPEG (Moving Picture Experts Group) system is the de facto standard as an encoding system for encoding moving images. In the MPEG system, SIMPLE, MAIN, SNR, SPATIAL, and HIGH are defined as profiles, and LOW, MAIN, HIGH 1440, and HIGH are defined as levels. MP @ ML (Main profile at main level) (hereinafter referred to as “SD”), which combines the main profile and the main level, corresponds to the current television system. On the other hand, MP @ HL (Main Profile at High Level) (hereinafter referred to as “HD”), which combines the main profile and the high level, corresponds to the HDTV system.

  Digital BS broadcasting, digital CS broadcasting, and terrestrial digital broadcasting have already begun. In these broadcasting, not only SD data but also HD data can be received at home. Therefore, a large screen monitor corresponding to HD data is gradually spreading.

  However, in order to record HD data, a large capacity recording medium is required. On the other hand, the hard disk drives that are currently on the market do not have a storage capacity for recording HD data for a long time. Also, the DVD (Digital Versatile Disc) that is currently popular does not have a storage capacity enough to record HD data for a long time.

  Therefore, in order to record a television program broadcast with HD data for a long time, it is necessary to convert the received HD data into SD data.

  In order to convert HD data to SD data, it is basically necessary to decode the HD data, reduce the number of pixels by some means, and then encode to SD data.

  Thus, a conversion device is disclosed that downsamples DCT coefficients obtained in the decoding process and passes the downsampled DCT coefficients to an encoding device (for example, Patent Documents 1 to 5).

Also, a conversion device is disclosed in which image data obtained by decoding is down-sampled in the horizontal and vertical directions and the image data after down-sampling is passed to an encoding device (for example, Patent Document 6).
JP-A-6-22291 JP-A-10-262252 JP 2001-285863 A JP 2001-285875 A JP 2002-34041 A JP 2000-78568 A

  However, according to the inventions described in Patent Documents 1 to 5, a special encoding device that can input DCT coefficients is required, and a general-purpose encoding device cannot be used.

  Further, the invention described in Patent Document 6 has the following problems.

  That is, in the HD data decoding device and the SD data encoding device, the components (luminance component (hereinafter referred to as “Y component”), first color difference component (hereinafter referred to as “U component”), and first. In order to collectively perform processing for each of two color difference components (hereinafter referred to as “V component”), a signal of a format called IYUV is used. On the other hand, in the interface between the HD data decoding device and the SD data encoding device, a signal of a format called YUY2 is used in consideration of compatibility with a display device and a capture device. Accordingly, the HD data decoding apparatus must perform conversion from the IYUV format to the YUY2 format. Also, the SD data encoding apparatus must perform reverse conversion from the YUY2 format to the IYUY format.

  Further, in the HD data decoding apparatus, since the IDCT (Inverse Discrete Cosine Transform) unit outputs the data in the word format, the HD data decoding apparatus must also perform conversion from the word format to the IYUV format. .

  Further, in the SD data encoding device, the DCT (Discrete Cosine Transform) unit inputs the data in the word format, and the IDCT unit outputs the data in the word format. In the previous stage, conversion from the IYUV format to the word format must be performed, and in the subsequent stage of the IDCT section, conversion from the word format to the IYUV format must be performed.

  Here, the data of each format will be described.

  First, the IYUV format data will be described with reference to FIG. 1. The ratio of the sampling interval of the Y component, the sampling interval of the U component, and the sampling interval of the V component is 1: 2: 2 in the horizontal direction, and the vertical direction. 1: 2: 2. Therefore, if the number of Y component samples per line is L, the Y component data Y0 to Y (2L-1) for two lines is included in the U component data U0 to U (L / 2 for one line). -1) corresponds to V component data V0 to V (L / 2-1) for one line. In order to perform processing for each component, the Y component data for one frame, the U component data for one frame, and the V component data for one frame are not mixed.

  Next, the data of the YUY2 format will be described with reference to FIG. 2. The ratio of the sampling interval of the Y component, the sampling interval of the U component, and the sampling interval of the V component is 1: 2: 2 in the horizontal direction, and is vertical. 1: 1: 1 in the direction. In order to perform display without rearranging data, the Y component, the U component, and the V component are mixedly arranged as shown in FIG.

  Next, the word format data will be described with reference to FIG. 3. The ratio of the sampling interval between the Y component, the U component and the V component and the ratio of the number of data are the same as in the IYUV format. One byte per byte is a byte having a value of 00H.

  Next, conversion from the word format to the IYUV format, conversion from the IYUV format to the word format, conversion from the IYUV format to the YUY2 format, and reverse conversion from the YUY2 format to the IYUV format are performed by Intel MMX (registered trademark) ( A case where an x86 processor having an instruction set of (Multi Media eXtension) is performed will be described.

  First, the conversion from the word format to the IYUV format will be described with reference to FIG. 4. Eight Y component samples are obtained by one packuswb instruction, and eight U component samples are obtained by one packuswb instruction. Eight V component samples are obtained by one packuswb instruction.

  In FIG. 4, one square represents one byte, and when one data is represented by a 16-bit word, the lower byte is written on the left and the upper byte is written on the right. The same applies to other similar drawings.

  Next, the conversion from the IYUV format to the word format will be described with reference to FIG. 5. Four Y component samples, four U component samples, and four V component samples are obtained by three punpcklbw instructions.

  Next, the conversion from the IYUV format to the YUY2 format will be described with reference to FIG. 6. Eight Y component samples, four U component samples, and four V component samples are obtained by three punpcklbw instructions.

  Next, with reference to FIG. 7 and FIG. 8, the inverse conversion from the YUY2 format to the IYUV format will be described. Eight Y component samples are obtained by two pand instructions, and by six psrlw instructions and two pand instructions. Eight U component samples and eight V component samples are obtained.

  As described above, the conversion from the word format to the IYUV format, the conversion from the IYUV format to the YUY2 format, and the reverse conversion from the YUY2 format to the IYUV format require many instructions. Therefore, a significant portion of the total processing time for conversion from HD data to SD data was spent for these conversions. This is an obstacle to realizing real-time conversion from HD data to SD data. If conversion from HD data to SD data cannot be performed in real time, simultaneous recording while receiving broadcast of HD data becomes impossible.

  Therefore, the present invention provides an image encoding method, an image encoding device, an image encoding program, an image decoding method, an image decoding device, and an image decoding that can speed up the conversion from HD data to SD data. It is an object to provide an image conversion program, an image conversion method, an image conversion apparatus, and an image conversion program.

  According to the first aspect of the present invention, by decoding encoded image data, luminance data expressed in a word format, first color difference data expressed in a word format, and second color difference expressed in a word format A decoding step for obtaining data, a luminance data conversion step for converting luminance data represented in the word format into luminance data represented in a byte format, first color difference data represented in the word format, and the word format A color difference data conversion step of converting the second color difference data represented by the first color difference data represented in a byte format and the second color difference data represented in a byte format into alternately arranged data. An image decoding method is provided.

  In the image decoding method, the luminance data conversion step may include a step of selecting lower 8 bits of the luminance data represented in the word format as luminance data represented in the byte format.

  In the image decoding method, the color difference data converting step includes a step of shifting the second color difference data in the word format, the shifted second color difference data in the word format, and the first color difference data in the word format. The step of performing the logical OR operation for each bit may be provided.

  The image decoding method described above includes luminance data represented in the byte format, first color difference data represented in the byte format, and second color difference data represented in the byte format, which are alternately arranged. , Luminance data represented by the even numbered byte format, first color difference data represented by the byte format, luminance data represented by the odd numbered byte format, and second color difference data represented by the byte format In this case, a luminance data / color difference data conversion step for converting the data into those arranged in this order may be further provided.

  In the image decoding method, the luminance data expressed in the byte format, the first color difference data expressed in the byte format, and the second color difference data expressed in the byte format, which are alternately arranged, The first image is represented, and the image decoding method is represented by luminance data represented by the byte format representing the first image, first color difference data represented by the byte format, and the byte format. Luminance data represented by the byte format and second color difference data represented by the byte format, and second color difference data representing a second image having a smaller number of pixels than the first image, and the first color difference data represented by the byte format, An image reduction step of converting the second color difference data expressed in the byte format into alternately arranged data may be further provided.

  According to the second aspect of the present invention, a luminance data reverse conversion step for reversely converting luminance data expressed in byte format into luminance data expressed in word format, first color difference data expressed in byte format, Color difference data reverse conversion step for reversely converting second color difference data expressed in byte format and alternately arranged into first color difference data expressed in word format and second color difference data expressed in word format; An encoding step of generating encoded image data from the luminance data expressed in the word format, the first color difference data expressed in the word format, and the second color difference data expressed in the word format. An image encoding method is provided.

  In the image coding method, the luminance inverse conversion step may include a step of generating luminance data represented in the word format by adding a bit string of zero to the luminance data represented in the byte format. Good.

  In the above image encoding method, the color difference data reverse conversion step includes the first color difference data expressed in the byte format and the second color difference data expressed in the byte format, which are alternately arranged and a predetermined mask. Obtaining a first color difference data represented in the word format by performing a logical product operation with a pattern for each bit; a first color difference data represented in the byte format; and a second color difference represented in the byte format. Obtaining the second color difference data expressed in the word format by shifting a predetermined number of bits of data arranged alternately.

  The image encoding method includes the even-numbered luminance data represented by the byte format, the first color difference data represented by the byte format, the odd-numbered luminance data represented by the byte format, and the byte format. The second color difference data expressed in this order, the luminance data expressed in the byte format, the first color difference data expressed in the byte format, and the second color difference data expressed in the byte format In addition, a luminance data / color difference data reverse conversion step may be further provided for reverse conversion to alternately arranged data.

  The image encoding method includes a luminance data conversion step of converting luminance data represented in the word format into luminance data represented in a byte format, and the first color difference data and the word format represented in the word format. A color difference data conversion step for converting the second color difference data represented by the above into the first color difference data represented in a byte format and the second color difference data represented in a byte format, which are alternately arranged, and the luminance Luminance data represented in the byte format obtained by the data conversion step, first color difference data represented in the byte format obtained in the color difference data conversion step, and second color difference data represented in the byte format. And a local decoding step of performing local decoding of the image data based on the alternately arranged data.

  In the image encoding method, the luminance data conversion step may include a step of selecting lower 8 bits of the luminance data represented in the word format as luminance data represented in the byte format.

  In the image encoding method, the color difference data conversion step includes a step of shifting the second color difference data in the word format, the shifted second color difference data in the word format, and the first color difference data in the word format. The step of performing the logical OR operation for each bit may be provided.

  According to a third aspect of the present invention, the image decoding method according to any one of claims 1 to 5 and the image encoding method according to any one of claims 6 to 12 are provided. A featured image conversion method is provided.

  According to the fourth aspect of the present invention, the spatial region data in the image decoding and the spatial region data in the image encoding have a continuous sample for the luminance signal and the first chrominance signal for the chrominance signal. There is provided an image conversion method characterized in that samples and samples of the second color difference signal are alternately continuous.

  According to the present invention, since the number of instructions required for conversion can be reduced, the execution time of the image decoding method, the image encoding method, and the image conversion method can be reduced.

  The best mode for carrying out the present invention will be described below in detail with reference to the drawings.

  The image conversion apparatus according to the present embodiment includes an image decoding apparatus and an image encoding apparatus.

  FIG. 9 shows an image decoding apparatus according to an embodiment of the present invention.

  Referring to FIG. 9, an image decoding apparatus according to an embodiment of the present invention includes a variable length decoding unit 101, an inverse quantization unit 102, an IDCT unit 103, a packing unit 104, an addition unit 105, a rear frame memory 106, a front frame memory 106, and a front frame memory 106. A frame memory 107, a motion compensation unit 108, a switch 109, an image reduction unit 110, and a rearrangement unit 111 are provided.

  The image decoding apparatus can be realized by hardware, but can also be realized by a computer reading and executing a program for causing a computer to function as the image decoding apparatus.

  The following description will be given in the case where the image decoding apparatus is realized by a computer, particularly regarding processing using a register in the CPU. However, since the CPU is realized by programmable hardware, The present invention can also be applied to hardware having a configuration similar to that of a CPU for realizing instructions related to the invention, in a programmable or fixed manner.

  The variable length decoding unit 101 performs variable length decoding of the HD MPEG video code in the HD data, and outputs quantized DCT coefficients, motion vectors, and the like.

  The inverse quantization unit 102 receives the quantized DCT coefficient from the variable length decoding unit 101, inversely quantizes the DCT coefficient, and outputs the DCT coefficient. This DCT coefficient is expressed in word format data. The word format data has already been described with reference to FIG.

  The IDCT unit 103 receives DCT coefficients from the inverse quantization unit 102, performs inverse discrete cosine transform on the DCT coefficients, and outputs spatial domain data. This spatial domain data is also expressed in word format.

  The packing unit 104 converts the spatial area data expressed in the word format into the spatial area data in the Y / UV pack format according to the present invention.

  Here, the Y / UV pack format will be described with reference to FIG. 10. The ratio of the sampling interval of the Y component, the sampling interval of the U component, and the sampling interval of the V component is 1: 1: 2: 2 and 1: 2: 2 in the vertical direction. Therefore, if the number of Y component samples per line is L, the Y component data Y0 to Y (2L-1) for two lines is included in the U component data U0 to U (L / 2 for one line). -1) corresponds to V component data V0 to V (L / 2-1) for one line. However, in the IYUV format, the U component data for one frame and the V component data for one frame are not mixed, whereas in the Y / UV pack format, as shown in FIG. V component bytes are arranged alternately. As described above, even if the U component and the V component are mixed, there is no problem because image processing is performed in units of macroblocks.

  Referring to FIG. 11, the conversion from the word format to the Y / UV pack format performed by the packing unit 104 will be described. By one packuswb instruction, one psllw instruction, and one por instruction, Four U component samples and four V component samples are obtained.

  The name of the packuswb instruction is “Pack words into bytes with unsigned saturation”, and the hardware for realizing the packuswb instruction is as shown in FIG. In FIG. 12, D0 to D7 indicate first registers, S0 to S7 indicate second registers, and FF0 to FF7 indicate D type flip-flops. When the clock is input to the D-type flip-flop, the contents of D0, D2, D4, D6, S0, S2, S4, and S6 are stored in D0 to D7, respectively.

  The name of the psllw instruction is “Shift packed words left logical”, and the hardware for realizing the psllw instruction is as shown in FIG. In FIG. 13, X0 to X3 indicate first registers, BS0 to BS3 indicate barrel shifters, and FF0 to FF3 indicate D-type flip-flops. When the number of shift bits is input to the barrel shifter and the clock is input to the D-type flip-flop, the contents of X0, X1, X2, and X3 are shifted to the left by the number of bits and stored in X0, X1, X2, and X3, respectively. The Zeros are stored in the parts freed by the shift.

  The name of the por instruction is “Packed or”, and hardware for realizing the por instruction is as shown in FIG. In FIG. 14, D0 to D7 indicate first registers, S0 to S7 indicate second registers, OR indicates an OR gate, and FF0 to FF7 indicate D type flip-flops. When the clock is input to the D-type flip-flop, the logical sum of the contents of D0 and S0 to the contents of D7 and the logical sum of the contents of S7 are stored in D0 to D7, respectively.

  The adding unit 105 adds the space area data supplied from the pack unit 104 to the space area data supplied from the switch 109. The output of the adding unit 105 follows the Y / UV pack format.

  The rear frame memory 106 and the previous frame memory 107 store I-picture spatial area data or P-picture spatial area data.

  The motion compensation unit 108 spatially moves the spatial area data read from the subsequent frame memory 106 or the previous frame memory 107 in units of macroblocks according to the motion vector supplied from the variable length decoding unit 101.

  The switch 109 selects zero when decoding an I picture, and selects the spatial domain data supplied from the motion compensation unit 108 when decoding a P picture or a B picture.

  The image reduction unit 110 downsamples the Y / UV pack format spatial region data supplied from the addition unit 105 from, for example, 1280 × 720 pixels to 720 × 480 pixels. The spatial area data output from the image reduction unit 110 also follows the Y / UV pack format.

  The rearrangement unit 111 converts the space area data in the Y / UV pack format into space area data in the YUY2 format. The YUY2 format is as described with reference to FIG.

  Referring to FIG. 15, the conversion from the Y / UV pack format to the YUY2 format performed by the rearrangement unit 111 will be described. By using two punpcklbw instructions, eight Y component samples, four U component samples, and four V components A sample is obtained.

  The name of the punpcklbw instruction is “Unpack low-order bytes”, and hardware for realizing the punpcklbw instruction is as shown in FIG. In FIG. 16, D0 to D7 indicate first registers, S0 to S7 indicate second registers, and FF0 to FF7 indicate D type flip-flops. When the clock is input to the D-type flip-flop, the contents of D0, D1, D2, D3, S0, S1, S2, and S3 are stored in D0 to D7, respectively.

  Since the space area data in the YUY2 format output from the rearrangement unit 111 is compatible with conventional devices, it can be supplied to conventional monitors and image encoding devices.

  FIG. 17 shows an image encoding device according to an embodiment of the present invention.

  Referring to FIG. 17, the image encoding apparatus according to the embodiment of the present invention includes an inverse rearrangement unit 141, a selection unit 142, a subtracter 143, an unpacking unit 144, a DCT unit 145, a quantization unit 146, and a variable length encoding unit. 147, an inverse quantization unit 148, an IDCT unit 149, a pack unit 150, an adder 151, a rear frame memory 152, a front frame memory 153, a motion search unit 154, a motion compensation unit 155, and a switch 156.

  The image encoding device can be realized by hardware, but can also be realized by a computer reading and executing a program for causing a computer to function as the image encoding device.

  The following description will be given in the case where the image encoding device is realized by a computer, particularly regarding processing using a register in the CPU. However, since the CPU is realized by programmable hardware, The present invention can also be applied to hardware having a configuration similar to that of a CPU for realizing instructions related to the invention, in a programmable or fixed manner.

  The reverse rearrangement unit 141 reversely converts the space area data in the YUY2 format into a space area in the Y / UV pack format.

  Referring to FIG. 18, the conversion from the YUY2 format to the Y / UV pack format performed by the reverse rearrangement unit 141 will be described. Eight Y component samples are obtained by two pand instructions, four psrlw instructions, and three packuswb instructions. Four U component samples and four V component samples are obtained.

  The name of the pslrw instruction is “Shift packed words right logical”, and the hardware for realizing the pslrw instruction is as shown in FIG. In FIG. 13, X0 to X3 indicate first registers, BS0 to BS3 indicate barrel shifters, and FF0 to FF3 indicate D-type flip-flops. When the number of shift bits is input to the barrel shifter and the clock is input to the D-type flip-flop, the contents of X0, X1, X2, and X3 are shifted to the right by the number of bits and stored in X0, X1, X2, and X3, respectively. The Zeros are stored in the parts freed by the shift.

  The name of the pand instruction is “Packed and”, and hardware for realizing the pand instruction is as shown in FIG. In FIG. 19, D0 to D7 indicate first registers, S0 to S7 indicate second registers, AND indicates an AND gate, and FF0 to FF7 indicate D type flip-flops. When the clock is input to the D-type flip-flop, the logical product of the content of D0 and the content of S0 to the content of D7 and the logical product of the content of S7 are stored in D0 to D7, respectively.

  Since the space area data in the YUY2 format input by the reverse rearrangement unit 141 is compatible with conventional devices, it can be input from a conventional capture device or image decoding device.

  The selection unit 142 selects the Y / UV pack format space area data supplied from the input interface or the Y / UV pack format space area data supplied from the reverse rearrangement unit 141.

  When the image decoding apparatus shown in FIG. 9 and the image encoding apparatus shown in FIG. 17 are connected, space area data can be supplied from the image decoding apparatus to the image encoding apparatus in the Y / UV pack format. The rearrangement unit 111 and the reverse rearrangement unit 141 are not necessary.

  The subtractor 143 subtracts the space area data supplied from the switch 156 in the Y / UV pack format from the Y / UV pack format space area data supplied from the selection unit 142.

  The unpack unit 144 receives Y / UV pack format space area data from the subtractor 143 and converts it into word format space area data.

  Referring to FIG. 20, the conversion from the Y / UV pack format to the word format performed by the unpacking unit 144 will be described. Four Y component samples are obtained by one punpcklbw instruction, one pand instruction, and one psrlw instruction. Four U component samples and four V component samples are obtained.

  The name of the punpckhbw instruction is “Unpack how-order bytes”, and hardware for realizing the punpckhbw instruction is as shown in FIG. In FIG. 21, D0 to D7 represent first registers, S0 to S7 represent second registers, and FF0 to FF7 represent D type flip-flops. When the clock is input to the D-type flip-flop, the contents of D4, D5, D6, D7, S4, S5, S6, and S7 are stored in D0 to D7, respectively.

  Next, the operation of the image decoding apparatus shown in FIG. 9 will be described with reference to FIG. 9, FIG. 22, and FIG.

  Referring to FIG. 22, first, an interface format with the image encoding device is determined (step S201). When the image encoding device shown in FIG. 17 is a connection partner, the interface format is determined as the Y / UV pack format. If the other device is the connection partner, the interface format is determined to be the YUY2 pack format.

  Next, an HD MPEG video code is input (step S202).

  Next, one frame is expanded to generate first Y / UV pack format data (step S203). Details of step S203 will be described later.

  Next, the image reduction unit 110 reduces the image to generate second Y / UV pack format data (step S204).

  Next, it is determined whether or not the interface format determined in step S201 is a Y / UV pack format (step S205).

  If so (YES in step S205), the second Y / UV pack format data is supplied to the image encoding device (step S206).

  If not (NO in step S205), the rearrangement unit 111 generates YUY2 format data from the second Y / UV pack format data (step S207).

  Next, YUY2 format data is supplied to the image encoding device (step S208).

  Next, it is determined whether or not the processing for all frames has been completed (step S209). Otherwise (NO in step S209), the process returns to step S202, and processing for the next frame is started. If so (YES in step S209), the process ends.

  Next, details of step S203 will be described with reference to FIG.

  Referring to FIG. 23, first, the current picture type is determined (step S301). If the current picture is an I picture, the process proceeds to step S302. If the current picture is a P picture, the process proceeds to step S308. If the current picture is a B picture, the process proceeds to step S318.

  In step S302, the variable length decoding unit 101 performs variable length decoding, and outputs quantized DCT coefficients and motion vectors.

  Next, the inverse quantization unit 102 inversely quantizes the quantized DCT coefficient and outputs the DCT coefficient (step S303).

  Next, the IDCT unit 103 performs inverse discrete cosine transform on the DCT coefficient, and outputs word format spatial domain data (step S304).

  Next, the packing unit 104 converts the space area data in the word format into the space area data in the Y / UV pack format (step S305).

  Next, space area data in the Y / UV pack format is written into the rear frame memory 106 (step S306).

  Next, the Y / UV pack format space area data is output to the image reduction unit 110 (step S307).

  In step S308, the variable length decoding unit 101 performs variable length decoding, and outputs quantized DCT coefficients and motion vectors.

  Next, the inverse quantization unit 102 inversely quantizes the quantized DCT coefficient and outputs the DCT coefficient (step S309).

  Next, the IDCT unit 103 performs inverse discrete cosine transform on the DCT coefficients and outputs word-format spatial domain data (step S310).

  Next, the packing unit 104 converts the space area data in the word format into space area data in the Y / UV pack format (step S311).

  Next, the contents of the rear frame memory 106 are transferred to the previous frame memory 107 (step S312). This transfer may be realized by switching the previous frame and the subsequent frame by operating the reference pointer.

  Next, the Y / UV pack format space area data of the reference frame is read from the previous frame memory 107 (step S313).

  Next, the motion compensation unit 108 spatially moves the reference frame data read out in step S313 in units of macroblocks according to the motion vector input from the variable length decoding unit 101 (step S314).

  Next, the adding unit 105 adds the spatial region data input from the packing unit 104 to the spatial region data input from the motion compensation unit 108 via the switch 109 (step S315).

  Next, the decoded Y / UV pack format data obtained by the addition in step S315 is written in the post-frame memory 106 (step S316) and output to the image reduction unit (step S317).

  Steps S318 to S327 are the same as steps S308 to S317, but since processing relating to the B picture is performed, in step S313 during processing relating to the P picture, the spatial area data in the Y / UV pack format of the reference frame is stored in the previous frame memory 107. On the other hand, in step S323 corresponding to this, the space area data in the Y / UV pack format of the reference frame is read from the previous frame memory 107 and the subsequent frame memory 106.

  Next, the operation of the image encoding device shown in FIG. 17 will be described with reference to FIGS. 17, 24, and 25.

  Referring to FIG. 24, first, an interface format with the image decoding apparatus is determined (step S221). When the image decoding apparatus shown in FIG. 9 is a connection partner, the interface format is determined as the Y / UV pack format. If the other device is the connection partner, the interface format is determined to be the YUY2 pack format.

  Next, it is determined whether or not the interface format determined in step S221 is a Y / UV pack format (step S222).

  If so (YES in step S222), the selection unit 142 inputs the second Y / UV pack format data from the image decoding apparatus (step S223).

  If not (NO in step S222), the reverse rearrangement unit 141 inputs YUY2 format data from the image decoding apparatus (step S224), and converts this into third Y / UV pack format data (step S225). ).

  Next, the current one frame is compressed to generate an SD MPEG video code (step S226). Details of step S226 will be described later.

  Next, the SD MPEG video code generated in step S226 is output (step S227).

  Next, it is determined whether or not the processing for all frames has been completed (step S228). Otherwise (NO in step S228), the process returns to step S222, and the process for the next frame is started. If so (YES in step S2228), the process ends.

  Next, details of step S226 will be described with reference to FIG.

  Referring to FIG. 25, first, the current picture type is determined (step S401). If the current picture is an I picture, the process proceeds to step S402. If the current picture is a P picture, the process proceeds to step S411. If the current picture is a B picture, the process proceeds to step S426.

  In step S402, the unpacking unit 144 converts the Y / UV pack format data input from the selection unit 142 via the subtracter 143 into word format data (step S402).

  Next, the DCT unit 145 subjects the spatial domain data in the word format input from the unpacking unit 144 to discrete cosine transform, and outputs DCT coefficients (step S403).

  Next, the quantization unit 146 quantizes the DCT coefficient input from the DCT unit 145, and outputs the quantized DCT coefficient (step S404).

  Next, the variable length coding unit 147 performs variable length coding on the quantized DCT coefficient input from the quantization unit 146 (step S405).

  Next, the inverse quantization unit 148 performs inverse quantization on the quantized DCT coefficient input from the quantization unit 146 and outputs a DCT coefficient in word format (step S406).

  Next, the IDCT unit 149 performs inverse discrete cosine transform on the DCT coefficients in the word format input from the inverse quantization unit 148, and outputs word format space region data (step S407).

  Next, the packing unit 150 converts the word-format space area data input from the inverse quantization unit 148 into Y / UV pack-format space area data (step S408).

  Next, the space area data in Y / UV pack format output by the packing unit 150 is written in the post-frame memory 152 (step S409).

  Next, the variable length code generated in step S405 is output as an SD MPEG video code (step S410).

  In step S411, the contents of the subsequent frame memory 152 are transferred to the previous frame memory 153. This transfer may be realized by switching the previous frame and the subsequent frame by operating the reference pointer.

  Next, the motion search unit 154 performs a motion vector for each macroblock based on the Y / UV pack format spatial region data input from the selection unit 142 and the Y / UV pack format spatial region data input from the previous frame memory 153. Is searched (step S412).

  Next, the Y / UV pack format space area data of the reference frame is read from the previous frame memory 153 (step S413).

  Next, the motion compensation unit 155 spatially moves the reference frame data read in step S413 in units of macroblocks according to the motion vector input from the motion search unit 154 (step S414).

  Next, the subtracter 143 subtracts the Y / UV pack format spatial domain data of the reference frame input from the motion compensation unit 155 via the switch 156 from the Y / UV pack format spatial domain data input from the selection unit 142. (Step S415).

  Next, the unpacking unit 144 converts the Y / UV pack format space area difference data input from the subtractor 143 into word format space area difference data (step S416).

  Next, the DCT unit 145 performs discrete cosine transform on the word-format space domain difference data input from the unpack unit 144, and outputs a word-format DCT coefficient (step S417).

  Next, the quantization unit 146 quantizes the DCT coefficient input from the DCT unit 145, and outputs the quantized DCT coefficient (step S418).

  Next, the variable length coding unit 147 performs variable length coding on the quantized DCT coefficient input from the quantization unit 146 (step S419).

  Next, the inverse quantization unit 148 inversely quantizes the quantized DCT coefficient input from the quantization unit 146, and outputs a DCT coefficient in word format (step S4420).

  Next, the IDCT unit 149 performs inverse discrete cosine transform on the DCT coefficients in the word format input from the inverse quantization unit 148, and outputs word format space region data (step S421).

  Next, the packing unit 150 converts the word-format space area data input from the inverse quantization unit 148 into Y / UV pack format space-area difference data (step S422).

  Next, the adder 151 adds the Y / UV pack format space area difference data input from the packing section 150 to the Y / UV pack format space area data of the motion compensated reference frame input from the motion compensation section 155. Addition is performed to generate space area restoration data of the current frame (step S423).

  Next, the space area restoration data in the Y / UV pack format obtained by the addition in step S423 is written in the rear frame memory 152 (step S424).

  Next, the variable length code generated in step S419 is output as an SD MPEG video code (step S425).

  In step S426, the contents of the subsequent frame memory 152 are transferred to the previous frame memory 153. This transfer may be realized by switching the previous frame and the subsequent frame by operating the reference pointer.

  Next, the motion search unit 154 receives Y / UV pack format space area data input from the selection unit 142, Y / UV pack format space area data input from the previous frame memory 153, and Y input from the back frame memory 152. A motion vector is searched for each macro block based on the spatial region data in the / UV pack format (step S427).

  Next, the Y / UV pack format space area data of the reference frame is read out from the previous frame memory 153 and the subsequent frame memory 152 (step S428).

  Next, the motion compensation unit 155 spatially moves the reference frame data read in step S428 in units of macroblocks according to the motion vector input from the motion search unit 154 (step S429).

  Next, the subtracter 143 subtracts the Y / UV pack format spatial domain data of the reference frame input from the motion compensation unit 155 via the switch 156 from the Y / UV pack format spatial domain data input from the selection unit 142. (Step S430).

  Next, the unpacking unit 144 converts the Y / UV pack format space area difference data input from the subtractor 143 into word format space area difference data (step S431).

  Next, the DCT unit 145 performs discrete cosine transform on the word-format space domain difference data input from the unpack unit 144, and outputs a word-format DCT coefficient (step S432).

  Next, the quantization unit 146 quantizes the DCT coefficient input from the DCT unit 145, and outputs the quantized DCT coefficient (step S433).

  Next, the variable length coding unit 147 performs variable length coding on the quantized DCT coefficient input from the quantization unit 146 (step S434).

  Next, the variable length code generated in step S434 is output as an SD MPEG video code (step S435).

  Next, an effect obtained by using Y / UV pack format data instead of IYUV format data will be described.

  In the space area of one macroblock, 256 (= (8 × 2) ^ 2) Y component samples, 64 (= 8 ^ 2) U component samples, 64 (= 8 ^ 2) There are V component samples.

  First, the conversion from the word format to the IYUV format performed by the conventional packing unit and the conversion from the word format to the Y / UV pack format performed by the packing unit 104 and the packing unit 150 of the present invention will be compared.

  According to the conventional example, as is apparent from FIG. 4, in order to perform conversion from the word format to the IYUV format within the range of one macroblock, 32 (= 256/8) packuswb for the Y component samples. For the instruction and U component sample, 8 (= 64/8) packuswb instructions are required, and for the V component sample, 8 (= 64/8) packuswb instructions are required. Therefore, in order to perform conversion from the word format to the IYUV format within the range of one macroblock, 48 (= 32 + 8 + 8) instructions are required for all components.

  On the other hand, according to the present invention, as is apparent from FIG. 11, in order to perform conversion from the word format to the Y / UV pack format within the range of one macroblock, 32 Y component samples (= 256 / Eight (8) packuswb instructions, U component and V component require 16 (= 64/4) psllw instructions and 16 (= 64/4) por instructions. Therefore, in order to perform conversion from the word format to the Y / UV pack format within the range of one macroblock, 64 (= 32 + 16 + 16) instructions are required for all components.

  Therefore, in contrast to the conversion from the word format to the IYUV format performed by the conventional packing unit, the conversion from the word format to the Y / UV packed format performed by the packing unit 104 and the packing unit 150 of the present invention is as follows. It requires 16 (= 64-48) more instructions.

  Next, the conversion from the IYUV format to the word format performed by the conventional unpacking unit and the conversion from the Y / UV pack format to the word format performed by the unpacking unit 144 of the present invention will be compared.

  According to the conventional example, as is apparent from FIG. 5, in order to perform conversion from the IYUV format to the word format within the range of one macroblock, 64 (= 256/4) punpcklbw are required for the Y component samples. For the U component sample, 16 (= 64/4) punpcklbw instructions are required, and for the V component sample, 16 (= 64/4) punpcklbw instructions are required. Therefore, in order to perform conversion from the IYUV format to the word format within the range of one macroblock, 96 instructions (= 64 + 16 + 16) are required for all components.

  On the other hand, according to the present invention, as is apparent from FIG. 20, in order to perform conversion from the Y / UV pack format to the word format within the range of one macroblock, 64 Y component samples (= 256 / 4) punpcklbw instructions, 16 (= 64/4) pand instructions for the U component, and 16 (= 64/4) psrlw instructions for the V component are required. Therefore, in order to perform conversion from the Y / UV pack format to the word format within the range of one macroblock, 96 (= 64 + 16 + 16) instructions are required for all components.

  Therefore, the conversion from the YYUV format to the word format performed by the unpacking unit 144 of the present invention requires the same number of instructions as compared to the conversion from the IYUV format to the word format performed by the conventional unpacking unit. And

  Next, the conversion from the IYUV format to the YUY2 format performed by the conventional rearrangement unit and the conversion from the Y / UV pack format to the YUY2 format performed by the rearrangement unit 111 of the present invention are compared.

  In the YUY2 format, one macroblock includes 256 Y component samples, 128 U component samples, and 128 V component samples.

  According to the conventional example, as is apparent from FIG. 6, in order to convert all components from the IYUV format to the YUY2 format within the range of one macroblock, 96 (= 3 × 256/8) punpcklbw Instruction is needed.

  On the other hand, according to the present invention, as is apparent from FIG. 15, in order to convert all components from the Y / UV pack format to the YUY2 format within the range of one macroblock, 64 (= 2 × 256 / 8) punpcklbw instructions are required.

  Therefore, in contrast to the conversion from the IYUV format to the YUY2 format performed by the conventional rearrangement unit, the conversion from the Y / UV pack format to the YUY2 format performed by the rearrangement unit 111 of the present invention is 32 (= 96). -64) Requires fewer instructions.

  Next, the conversion from the YUY2 format to the IYUV format performed by the conventional reverse rearrangement unit and the conversion from the YUY2 format to the Y / UV pack format performed by the reverse rearrangement unit 141 of the present invention are compared.

  According to the conventional example, as is apparent from FIGS. 7 and 8, in order to perform conversion from the YUY2 format to the IYUV format within the range of one macroblock, 64 Y component samples (= 2 × 256 / 8) pand instructions and 32 (= 256/8) packuswb instructions, and 48 (= 6 × 64/8) psrlw instructions for U component samples and V component samples, 32 (= 4 ×) 64/8) packuswb instructions and 16 (= 2 × 64/8) pand instructions are required. Therefore, in order to perform conversion from the YUY2 format to the IYUV format within the range of one macroblock, 192 (= (64 + 32) + (48 + 32 + 16)) instructions are required for all components.

  On the other hand, according to the present invention, as is apparent from FIG. 18, in order to perform conversion from the YUY2 format to the Y / UV pack format within the range of one macroblock, 64 Y component samples (= 2 ×) 256/8) pand instructions and 32 (= 256/8) packuswb instructions. For U component samples and V component samples, 32 (= 4 × 64/8) psrlw instructions and 16 (= 2 × 64/8) packuswb instructions are required. Accordingly, in order to perform conversion from the YUY2 format to the Y / UV pack format within the range of one macroblock, 144 instructions (= (64 + 32) + (32 + 16)) are required for all components.

  Therefore, 48 conversions from the YUY2 format to the Y / UV pack format performed by the reverse sorting unit 141 of the present invention are performed in contrast to the conversion from the YUY2 format to the IYUV format performed by the conventional reverse sorting unit. = 192-144) Requires fewer instructions.

  The above comparison is summarized in Table 1.

また、ＨＤの画面サイズを１２８０×７２０画素、ＳＤの画面サイズを７２０×４８０画素とすると、両者の画素数の比率は１：０．３７５となる。そして、画像復号化装置は、パック化部及び並び替え部を備えるが、アンパック化部及び逆並び替え部を備えない。また、画像符号化装置はパック化部、アンパック化部及び逆並び替え部を備えるが、並び替え部を備えない。そうすると、命令数の重み付けは表２に示すとおりとなる。表１における重み１は表２における合計と等しい。

If the HD screen size is 1280 × 720 pixels and the SD screen size is 720 × 480 pixels, the ratio of the number of pixels is 1: 0.375. The image decoding apparatus includes a packing unit and a rearranging unit, but does not include an unpacking unit and a reverse rearranging unit. Further, the image coding apparatus includes a packing unit, an unpacking unit, and a reverse rearrangement unit, but does not include a rearrangement unit. Then, the weighting of the number of instructions is as shown in Table 2. The weight 1 in Table 1 is equal to the sum in Table 2.

また、ＨＤからＳＤへの変換を行わない場合には、命令数の重み付けは表３に示すとおりとなる。表１における重み２は表３における合計と等しい。

When conversion from HD to SD is not performed, the weighting of the number of instructions is as shown in Table 3. The weight 2 in Table 1 is equal to the sum in Table 3.

表１の最下行に示すように、単純に増減数を比較すると、本発明による命令数は、従来例による命令数よりも１５％少ない。また、重み１を考慮すると、本発明による命令数は、従来例による命令数よりも６％少ない。更に、重み２を考慮すると、本発明による命令数は、従来例による命令数よりも１１％少ない。

As shown in the bottom row of Table 1, when the number of increase / decrease is simply compared, the number of instructions according to the present invention is 15% less than the number of instructions according to the conventional example. Also, considering the weight 1, the number of instructions according to the present invention is 6% less than the number of instructions according to the conventional example. Further, considering the weight 2, the number of instructions according to the present invention is 11% less than the number of instructions according to the conventional example.

  The present invention can be used to decode image data, to encode image data, and to convert image data.

It is a figure which shows the image data represented by an IYUV format. It is a figure which shows the image data represented by a YUY2 format. It is a figure which shows the image data represented by a word format. It is a figure which shows conversion from a word format to an IYUV format. It is a figure which shows conversion from an IYUV format to a word format. It is a figure which shows the conversion from an IYUV format to a YUY2 format. It is a 1st figure which shows the conversion from a YUY2 format to an IYUV format. It is a 2nd figure which shows the conversion from a YUY2 format to an IYUV format. It is a block diagram which shows the structure of the image decoding apparatus by embodiment of this invention. It is a figure which shows the image data represented by the Y / UV pack format by embodiment of this invention. It is a figure which shows conversion from the word format by the embodiment of this invention to the Y / UV pack format. It is a figure which shows the circuit which implement | achieves a packuswb instruction. It is a figure which shows the circuit which implement | achieves a psrlw instruction and a psllw instruction. It is a figure which shows the circuit which implement | achieves a por instruction. It is a figure which shows the conversion from Y / UV pack format to YUY2 format by embodiment of this invention. It is a figure which shows the circuit which implement | achieves a punpcklbw instruction. It is a block diagram which shows the structure of the image coding apparatus by embodiment of this invention. It is a figure which shows the conversion from the YUY2 format to the Y / UV pack format by embodiment of this invention. It is a figure which shows the circuit which implement | achieves a pand instruction | indication. It is a figure which shows conversion from the Y / UV pack format to the word format by embodiment of this invention. It is a figure which shows the circuit which implement | achieves a punpckhbw instruction. 5 is a flowchart illustrating an image decoding method according to an embodiment of the present invention. It is a flowchart which shows the frame expansion | extension method shown in FIG. 5 is a flowchart illustrating an image encoding method according to an embodiment of the present invention. It is a flowchart which shows the frame compression method shown in FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Variable length decoding part 102 Inverse quantization part 103 IDCT part 104 Packing part 105 Adder part 106 Later frame memory 107 Previous frame memory 108 Motion compensation part 109 Switch 110 Image reduction part 111 Rearrangement part 141 Reverse rearrangement part 142 Selection Unit 143 subtractor 144 unpack unit 145 DCT unit 146 quantization unit 147 variable length coding unit 148 inverse quantization unit 149 IDCT unit 150 pack unit 151 adder 152 subsequent frame memory 153 previous frame memory 154 motion search unit 155 motion compensation unit 156 switch

Claims (32)

Decoding step of obtaining luminance data represented in a word format, first color difference data represented in a word format, and second color difference data represented in a word format by decoding the encoded image data;
A luminance data conversion step for converting the luminance data represented in the word format into luminance data represented in a byte format;
The first color difference data expressed in the word format and the second color difference data expressed in the word format are first color difference data expressed in byte format and second color difference data expressed in byte format, alternately. A color difference data conversion step for converting to a lineup of
An image decoding method comprising: The image decoding method according to claim 1,
The luminance data conversion step includes
An image decoding method comprising: selecting lower 8 bits of luminance data expressed in the word format as luminance data expressed in the byte format. The image decoding method according to claim 1,
The color difference data conversion step includes
Shifting the second color difference data in the word format;
Performing a logical OR operation on the shifted second color difference data in the word format and the first color difference data in the word format for each bit;
An image decoding method comprising: The image decoding method according to claim 1,
The luminance data expressed in the byte format, the first color difference data expressed in the byte format, and the second color difference data expressed in the byte format are alternately arranged in the even-numbered byte format. Luminance data represented, first color difference data represented in the byte format, odd-numbered luminance data represented in the byte format, and second color difference data represented in the byte format, arranged in this order An image decoding method, further comprising: a luminance data / color difference data conversion step for converting into a color data. The image decoding method according to claim 1,
The luminance data expressed in the byte format, the first color difference data expressed in the byte format, and the second color difference data expressed in the byte format, which are alternately arranged, represent the first image. ,
The luminance data represented in the byte format representing the first image, the first color difference data represented in the byte format, and the second color difference data represented in the byte format, which are alternately arranged, The luminance data represented by the byte format representing the second image having a smaller number of pixels than one image, the first color difference data represented by the byte format, and the second color difference data represented by the byte format, alternately. An image decoding method, further comprising: an image reduction step for converting the image into a lineup. A luminance data reverse conversion step for reversely converting the luminance data represented in byte format into luminance data represented in word format;
The first color difference data expressed in byte format and the second color difference data expressed in word format and second color difference data expressed in byte format and second color difference data expressed in byte format and second color difference expressed in word format A color difference data reverse conversion step for reverse conversion to data;
An encoding step for generating encoded image data from the luminance data expressed in the word format, the first color difference data expressed in the word format, and the second color difference data expressed in the word format;
An image encoding method comprising: The image encoding method according to claim 6, wherein
The luminance inverse conversion step includes:
An image encoding method comprising: generating luminance data represented in the word format by adding a bit string of zero to the luminance data represented in the byte format. The image encoding method according to claim 6, wherein
The color difference data reverse conversion step includes:
The word format is obtained by performing a logical product operation between the first color difference data expressed in the byte format and the second color difference data expressed in the byte format alternately arranged and a predetermined mask pattern for each bit. Obtaining first color difference data represented by:
The first color difference data represented in the byte format and the second color difference data represented in the byte format, which are alternately arranged, are shifted by a predetermined number of bits to obtain the second color difference data represented in the word format. Obtaining step;
An image encoding method comprising: The image encoding method according to claim 6, wherein
The even-numbered luminance data represented by the byte format, the first color difference data represented by the byte format, the luminance data represented by the odd-numbered byte format, and the second color difference data represented by the byte format. What is arranged in this order is the luminance data expressed in the byte format, the first color difference data expressed in the byte format, and the second color difference data expressed in the byte format, which are alternately arranged. An image coding method, further comprising: a luminance data / color difference data reverse conversion step for performing reverse conversion to The image encoding method according to claim 6, wherein
A luminance data conversion step for converting the luminance data represented in the word format into luminance data represented in a byte format;
The first color difference data expressed in the word format and the second color difference data expressed in the word format are first color difference data expressed in the byte format and second color difference data expressed in the byte format, alternately. A color difference data conversion step for converting to a lineup of
The luminance data expressed in the byte format obtained in the luminance data conversion step, the first color difference data expressed in the byte format obtained in the color difference data conversion step, and the second color difference data expressed in the byte format. And a local decoding step for performing local decoding of the image data based on the alternating array,
An image encoding method, further comprising: The image encoding method according to claim 10, wherein
The luminance data conversion step includes
An image encoding method comprising: selecting lower 8 bits of luminance data represented in the word format as luminance data represented in the byte format. The image encoding method according to claim 10, wherein
The color difference data conversion step includes
Shifting the second color difference data in the word format;
Performing a logical OR operation on the shifted second color difference data in the word format and the first color difference data in the word format for each bit;
An image encoding method comprising:   An image conversion method comprising the image decoding method according to any one of claims 1 to 5 and the image encoding method according to any one of claims 6 to 12. Decoding means for obtaining luminance data represented in word format, first color difference data represented in word format, and second color difference data represented in word format by decoding encoded image data;
Luminance data conversion means for converting the luminance data represented in the word format into luminance data represented in a byte format;
The first color difference data expressed in the word format and the second color difference data expressed in the word format are first color difference data expressed in byte format and second color difference data expressed in byte format alternately. Color difference data conversion means for converting to a lineup of
An image decoding apparatus comprising: The image decoding device according to claim 14, wherein
The luminance data converting means includes
An image decoding apparatus comprising: means for selecting lower 8 bits of luminance data expressed in the word format as luminance data expressed in the byte format. The image decoding device according to claim 14, wherein
The color difference data converting means includes
Means for shifting the second color difference data in the word format;
Means for performing bitwise OR operation on the shifted second color difference data in the word format and the first color difference data in the word format;
An image decoding apparatus comprising: The image decoding device according to claim 14, wherein
The luminance data represented in the byte format, the first color difference data represented in the byte format, and the second color difference data represented in the byte format, which are alternately arranged in the even-numbered byte format. Luminance data represented, first color difference data represented in the byte format, odd-numbered luminance data represented in the byte format, and second color difference data represented in the byte format, arranged in this order An image decoding apparatus, further comprising: luminance data / color difference data conversion means for converting the image data into image data. The image decoding device according to claim 14, wherein
The luminance data expressed in the byte format, the first color difference data expressed in the byte format, and the second color difference data expressed in the byte format, which are alternately arranged, represent the first image. ,
The luminance data represented in the byte format representing the first image, the first color difference data represented in the byte format, and the second color difference data represented in the byte format, which are alternately arranged, The luminance data represented by the byte format representing the second image having a smaller number of pixels than one image, the first color difference data represented by the byte format, and the second color difference data represented by the byte format, alternately. An image decoding apparatus, further comprising: an image reduction means for converting the image data to those arranged in line. Luminance data reverse conversion means for reversely converting luminance data represented in byte format into luminance data represented in word format;
The first color difference data expressed in byte format and the second color difference data expressed in word format and second color difference data expressed in byte format and second color difference data expressed in byte format and second color difference expressed in word format Color difference data reverse conversion means for performing reverse conversion to data;
Encoding means for generating encoded image data from the luminance data expressed in the word format, the first color difference data expressed in the word format, and the second color difference data expressed in the word format;
An image encoding device comprising: The image encoding device according to claim 19,
The luminance inverse conversion means includes
An image encoding apparatus comprising: means for generating luminance data expressed in the word format by adding a bit string of zero to the luminance data expressed in the byte format. The image encoding device according to claim 19,
The color difference data reverse conversion means includes:
The word format is obtained by performing a logical product operation between the first color difference data expressed in the byte format and the second color difference data expressed in the byte format alternately arranged and a predetermined mask pattern for each bit. Means for obtaining first color difference data represented by:
The first color difference data represented in the byte format and the second color difference data represented in the byte format, which are alternately arranged, are shifted by a predetermined number of bits to obtain the second color difference data represented in the word format. Means to obtain,
An image encoding device comprising: The image encoding device according to claim 19,
The even-numbered luminance data represented by the byte format, the first color difference data represented by the byte format, the luminance data represented by the odd-numbered byte format, and the second color difference data represented by the byte format. What is arranged in this order is the luminance data expressed in the byte format, the first color difference data expressed in the byte format, and the second color difference data expressed in the byte format, which are alternately arranged. An image encoding apparatus, further comprising: luminance data / color difference data reverse conversion means for performing reverse conversion to the above. The image encoding device according to claim 19,
Luminance data conversion means for converting the luminance data represented in the word format into luminance data represented in a byte format;
The first color difference data expressed in the word format and the second color difference data expressed in the word format are first color difference data expressed in the byte format and second color difference data expressed in the byte format, alternately. Color difference data conversion means for converting to a lineup of
The luminance data represented by the byte format obtained by the luminance data converting means, the first color difference data represented by the byte format obtained by the color difference data converting means, and the second color difference data represented by the byte format. And local decoding means for performing local decoding of the image data based on the alternating array,
An image encoding apparatus, further comprising: The image encoding device according to claim 23, wherein
The luminance data converting means includes
An image encoding apparatus comprising: means for selecting lower 8 bits of luminance data expressed in the word format as luminance data expressed in the byte format. The image encoding device according to claim 23, wherein
The color difference data converting means includes
Means for shifting the second color difference data in the word format;
Means for performing bitwise OR operation on the shifted second color difference data in the word format and the first color difference data in the word format;
An image encoding device comprising:   An image conversion apparatus comprising: the image decoding apparatus according to any one of claims 14 to 18; and the image encoding apparatus according to any one of claims 19 to 25.   The program for making a computer perform the image decoding method of any one of Claims 1 thru | or 5.   The program for making a computer perform the image coding method of any one of Claims 6 thru | or 12.   A program for causing a computer to perform the image conversion method according to claim 13.   The format of the spatial domain data in the image decoding and the spatial domain data in the image coding is such that the samples for the luminance signal are continuous, and for the color difference signal, the first color difference signal sample and the second color difference signal sample are alternated. An image conversion method characterized in that it is continuous.   The format of the spatial domain data in the image decoding and the spatial domain data in the image coding is such that the samples for the luminance signal are continuous, and for the color difference signal, the first color difference signal sample and the second color difference signal sample are alternated. An image conversion device characterized in that the image conversion device is continuous.   The format of the spatial domain data in the image decoding and the spatial domain data in the image coding is such that the samples for the luminance signal are continuous, and for the color difference signal, the first color difference signal sample and the second color difference signal sample are alternated. A program for causing a computer to perform an image conversion method characterized by being continuous to the above.

Images