JP2007020123A - Apparatus and method for detecting motion - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus and method for detecting motion in which the amount of data to be transferred to a local memory is reduced, seen by a plurality of encoding target pictures in image encoding processing. <P>SOLUTION: The motion detection apparatus comprises (a) an encoding target block local memory 111 for storing pixel data in a first encoding target image region, belonging to a first encoding target picture; (b) an encoding target block local memory 112 for storing pixel data in a second encoding target image region, belonging to a second encoding target picture; (c) a reference local memory 52 for storing pixel data in a reference target image region, belonging to a reference picture to be used for motion detection of the first encoding target picture and the second encoding target picture; (d) and a motion detector 113 for detecting motion between pictures, by performing block matching upon the pixel data in the reference target image region, the pixel data in the first encoding target image region, and the pixel data in the second encoding target image region, respectively. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a motion detection apparatus that detects a motion between pictures by performing a block matching operation on pixel data of an encoding target image region belonging to an encoding target picture and pixel data of a reference target image region belonging to a reference picture. About.

  In recent years, the multimedia era has come to handle voice, images, and other pixel values in an integrated manner, and traditional information media, that is, means for transmitting information such as newspapers, magazines, televisions, radios, and telephones to people are multimedia. It has come to be taken up as a target of. In general, multimedia refers to not only characters but also figures, sounds, especially images, etc. that are associated with each other at the same time. To make these information media multimedia, the information is converted into digital form. It is a necessary condition to express.

  However, when the information amount of each information medium is estimated as a digital information amount, the amount of information per character is 1 to 2 bytes in the case of characters, while 64 Kbit per second in the case of speech. (Phone quality) is required. Furthermore, for a moving image, an information amount of 100 Mbit per second (current television reception quality) or more is required. As described above, it is not realistic to handle such an enormous amount of information as it is in digital format. For example, a video phone has already been put into practical use by an integrated services digital network (ISDN) having a transmission rate of 64 Kbit / s to 1.5 Mbit / s. However, it is impossible to send the video of the TV camera as it is with ISDN.

  Therefore, information compression technology is required. For example, in the case of a videophone, H.264 recommended by ITU-T (International Telecommunication Union Telecommunication Standardization Sector). 261 and H.264. H.263 standard video compression technology is used. In addition, according to the information compression technology of the MPEG-1 standard, it is possible to put image information together with audio information on a normal music CD (compact disc).

  Here, MPEG (Moving Picture Experts Group) is an international standard for video signal compression standardized by ISO / IEC (International Electrotechnical Commission). This is a standard for compressing information of a television signal up to 5 Mbps, that is, about 1/100. In the MPEG-1 standard, the target quality is a medium quality, specifically, a quality that can be realized mainly at a transmission rate of about 1.5 Mbps. On the other hand, MPEG-2, which is standardized to meet the demand for higher image quality, realizes TV broadcast quality at 2 to 15 Mbps for moving image signals. Furthermore, at present, the work group (ISO / IEC JTC1 / SC29 / WG11) that has been standardizing with MPEG-1 and MPEG-2 achieves a compression ratio that exceeds MPEG-1 and MPEG-2, and further, unit by object. MPEG-4 has been standardized, which enables encoding, decoding, and operation and realizing new functions necessary in the multimedia era. In MPEG-4, it was originally aimed at standardizing a low bit rate encoding method, but now it has been extended to a more general encoding including a high bit rate including interlaced images.

  Further, in 2003, MPEG-4 AVC and ITU H.264 were jointly developed by ISO / IEC and ITU-T as image compression systems with higher compression rates. H.264 is standardized (for example, see Non-Patent Document 1). H. The H.264 standard is currently drafting an amended standard draft compatible with High Profile suitable for HD (High Definition) images and the like. H. Applications of the H.264 standard are expected to spread to digital broadcasting, DVD players / recorders, hard disk players / recorders, camcorders, videophones, and the like, similar to MPEG-2 and MPEG-4.

  In general, in encoding of moving images, the amount of information is compressed by reducing redundancy in the time direction and the spatial direction. Therefore, in inter-screen predictive coding for the purpose of reducing temporal redundancy, motion detection and prediction image creation are performed in block units with reference to the front or rear picture, and the obtained prediction image and code are obtained. Encoding is performed on the difference value from the current picture. Here, a picture is a term representing a single screen, which means a frame in a progressive image and a frame or field in an interlaced image. Here, an interlaced image is an image in which one frame is composed of two fields having different times. In interlaced image encoding and decoding processing, one frame may be processed as a frame, processed as two fields, or processed as a frame structure or a field structure for each block in the frame. it can.

  A picture that does not have a reference picture and performs intra prediction coding is called an I picture. A picture that performs inter-frame predictive coding with reference to only one reference picture is called a P picture. A picture that can be subjected to inter-picture prediction coding with reference to two reference pictures at the same time is called a B picture. The B picture can refer to two pictures as an arbitrary combination of display times from the front or the rear. A reference picture (reference picture) can be specified for each macroblock that is a basic unit of encoding. The reference picture described earlier in the encoded bitstream is the first reference picture, The one described is distinguished as the second reference picture. However, as a condition for encoding these pictures, the picture to be referenced needs to be already encoded. In addition, in the P picture and the B picture, it is possible to select the coding unit block in which the intra-frame prediction coding is performed or the inter-frame prediction coding is performed.

  Motion compensation inter-picture prediction coding is used for coding a P picture or a B picture. The motion compensation inter-picture prediction encoding is an encoding method in which motion compensation is applied to inter-picture prediction encoding. Motion compensation is not simply predicting from the pixel value of the reference frame, but detecting the amount of motion of each part in the picture (hereinafter referred to as a motion vector) and performing prediction in consideration of the amount of motion. This improves the prediction accuracy and reduces the amount of data. For example, the amount of data is reduced by detecting the motion vector of the encoding target picture and encoding the prediction residual between the prediction value shifted by the motion vector and the encoding target picture. In the case of this method, since motion vector information is required at the time of decoding, the motion vector is also encoded and recorded or transmitted.

  The motion vector is detected in units of macroblocks. Specifically, the macroblock on the encoding target picture side is fixed, the macroblock on the reference picture side is moved within the search range, and is most similar to the reference block. The motion vector is detected by finding the position of the reference block.

  FIG. 9 is a block diagram showing a configuration of a conventional inter-picture prediction encoding apparatus. As shown in the figure, the inter-picture prediction encoding apparatus 10 includes an input memory 11, a multi-frame memory 12, a motion detection unit 13, a motion compensation unit 14, a motion vector prediction / DifMV calculation unit 15, a motion vector memory 16, A frequency conversion / quantization unit 17, an inverse frequency conversion / inverse quantization unit 18, a regeneration unit 19, an encoding processing unit 20, and the like are provided.

  The motion detection unit 13 compares the motion detection reference pixel 31 output from the multi-frame memory 12 with the original image pixel 32 stored in the input memory 11, and outputs a motion vector 33 and a reference frame number 34. . The reference frame number 34 is an identification signal that identifies a reference image that is selected from a plurality of reference images and that is referred to by the target image. The motion vector MV is temporarily stored in the motion vector memory 16 and then output as a neighborhood motion vector 35. As a neighborhood motion vector referred to in order to predict a motion vector predictor by the motion vector prediction / DifMV calculation unit 15. used. The motion vector prediction / DifMV calculation unit 15 subtracts the predicted motion vector from the motion vector 33 and outputs the difference as a motion vector prediction difference 36.

  On the other hand, the multi-frame memory 12 outputs the pixel indicated by the reference frame number 34 and the motion vector 33 as the motion compensation reference pixel 37. The motion compensation unit 14 generates a reference pixel with decimal pixel accuracy and outputs a screen prediction error 38 with the original image 32 from the input memory 11.

  The screen prediction error 38 is subjected to frequency conversion and quantization processing by the frequency conversion / quantization unit 17, and the quantized coefficient 39 is output to the encoding processing unit 20 at the same time as the inverse frequency conversion / inverse quantization unit. At 18, the inverse quantization process and the inverse frequency conversion are performed. The post-quantization coefficient 39, the motion vector prediction difference 36, and the reference frame number 34 output from the frequency transform / quantization unit 17 are variable-length encoded by the encoding processing unit 20, and an encoded signal 42 is output. The code screen prediction error 40 output from the inverse frequency transform / inverse quantization unit 18 is added to the screen prediction error 38 in the regeneration unit 19 and is stored in the multiframe memory 12 as the code pixel 41. However, in order to make effective use of the capacity of the multi-frame memory 12, the screen area stored in the multi-frame memory 12 is released when it is unnecessary, and the decoded screen of the screen that does not need to be stored in the multi-frame memory 12 41 is not stored in the multi-frame memory 12.

  By the way, the structure for mounting the inter-screen prediction encoding apparatus shown in FIG. 9 is shown as an example in Patent Document 1, for example. As shown in Patent Document 1, when the inter-picture prediction encoding apparatus 10 is mounted on an LSI or the like, the multi-frame memory 12 shown in FIG. 9 is an external frame memory connected outside the LSI. And divided into a local memory inside the LSI that is directly accessed at the time of block matching search by the motion detection unit 13. The input memory 11 is also divided into an external input memory and an encoding target block local memory inside the LSI.

  FIG. 10 is a diagram for explaining a connection configuration between the inter-picture prediction encoding apparatus and the frame memory. In the figure, the same reference numerals as those in FIG. 9 have the same functions, and the description thereof is omitted. The external frame memory 51 is a frame memory realized outside the LSI. The reference local memory 52 is a local memory inside the LSI, and is directly accessed when a block matching search is performed by the motion detector 13 and when a motion compensated prediction pixel is generated by the motion compensator 14. The external input memory 53 is a frame memory realized outside the LSI, and stores an input image. The encoding target block local memory 54 is a local memory inside the LSI, and is directly accessed by the motion detector 13 and the motion compensator 14 at the time of block matching and when a screen prediction error is generated. The motion detection device 50 represents a portion mounted on the LSI of the inter-screen predictive coding device 10.

  When motion detection is performed in FIG. 10, first, an image area to be searched is transferred from the external frame memory 51 to the reference local memory 52 via the external bus 61. The original image to be encoded is transferred from the external input memory 53 to the encoding target block local memory 54 via the external bus 62. Next, data is read from the reference local memory 52 and the encoding target block local memory 54 via the internal buses 63 and 64, and motion detection is performed by the motion detector 13. When motion compensation is performed, the image area at the position determined by motion detection is transferred from the external frame memory 51 to the reference local memory 52 via the external bus 61, and the motion compensator 14 uses this to transfer motion. A compensation prediction pixel is generated, and a screen prediction error with the original image from the encoding target block local memory 54 is generated. By adopting such a configuration, the capacity of the internal memory of the LSI is reduced.

  FIG. 11 shows an example of the data structure of the reference local memory 52. 11A shows an MPEG-2 standard definition (SD) size image, and FIG. H.264 SD size image, FIG. 11C shows an MPEG-2 HD size image, and FIG. Each state of the local memory for reference when an H.264 HD size image is assumed is shown.

  FIG. 12 is a schematic diagram showing the state of reference pixels transferred per screen. In the figure, when motion detection processing is performed in units of 1 MB (Macro Block) (= 16 pixels × 16 pixels), in order to perform motion detection for 1 MB rows, (vertical length of search range) × (1 This means that memory transfer for the horizontal width of the screen is required. When motion detection for one screen is performed, the memory transfer amount for multiplying the number of vertical MBs on one screen is required. ing.

  For example, as shown in FIG. 11A, generally, transfer processing from the external frame memory 51 to the reference local memory 52 and motion detection processing performed using the reference local memory 52 and the motion detector 13 are performed. In order to perform parallel operation and to improve the overall operation speed, an update area indicated by hatching is also required. That is, in an SD size image (720 pixels × 480 pixels, 45 MB × 30 MB) such as MPEG-2, when the search range is from −16 to +16 in the vertical and horizontal directions, (16 + 16 × 2) × 720 × 30 = 1,036 800 pixels are transferred by motion detection per screen.

  As shown in FIG. In the management state of the reference local memory assuming an H.264 SD size image, H.264 is used. In H.264, since a 6-tap filter is used for motion compensation with decimal pixel accuracy, more peripheral pixels are required than in conventional MPEG-2. That is, in MPEG-2 or the like, a decimal precision pixel is created from four integer pixels surrounding a decimal precision pixel position. However, in the case of a 6-tap filter, decimal precision pixels are created from 36 integer pixels, so even when searching in the same region, compared to MPEG-2, the top 2 rows, the bottom 2 rows, the left 2 columns, the right 2 A column of pixels is required. Therefore, H.I. In an SD size image such as H.264, when the search range is −16 to +16 in the vertical and horizontal directions, (16 + 16 × 2 + 4) × 720 × 30 = 1, 123,200 pixels are transferred by motion detection per screen. Become.

  Further, as shown in FIG. 11C, in the management state of the reference local memory assuming an MPEG-2 HD size image, the HD size image has about six times the number of pixels as the SD size image. Therefore, for the sake of simplicity, the reference area is searched 2.5 times in the vertical and horizontal directions. Accordingly, the vertical and horizontal search range is −40 to +39 pixels, and similarly, (16 + 40 × 2) × 1,920 × 68 = 12,533,760 pixels are transferred by motion detection per screen. Become.

  Further, as shown in FIG. Similarly, in the management state of the reference local memory assuming an H.264 HD size image, (16 + 40 × 2 + 4) × 1,920 × 68 = 13,056,000 pixels are transferred by motion detection per screen. It will be. As a result, the transfer amount is an order of magnitude larger than the SD size of MPEG-2, and the power consumption is also increased.

  That is, when an HD size image (1920 pixels × 1088 pixels, 120 MB × 68 MB) is to be handled, the In the case of encoding with H.264, the pixel transfer amount per screen greatly increases, which may exceed the transfer capability of the external bus 61 shown in FIG.

On the other hand, a technique for reducing data transfer from an external frame memory to an internal reference local memory by sharing a motion vector search region among a plurality of coding blocks has been proposed (for example, Patent Document 2). reference.).
ISO / IEC 14496-10, International Standard: "Information technology-Coding of audio-visual objects-Part 10: Advanced video coding" (2003-12-01) Japanese Patent No. 2963269 JP 2003-134518 A

  However, in the conventional technique, when an image having a large size such as an HD image size is subjected to inter-frame predictive encoding, When performing inter-picture predictive encoding using H.264 or the like, the compression rate can be improved by detecting motions in a plurality of or a wide range of reference areas. However, the amount of memory transfer when transferring a reference pixel from an external memory to an internal logic is small. It becomes enormous and a problem arises in the bus transfer capability.

  On the other hand, for example, in the technique disclosed in Patent Document 2, it is possible to reduce the data transfer amount to the local memory when viewed in units of pictures to be encoded. However, this is only a reduction in the amount of data transfer in units of pictures. In other words, there is room for further reduction in the amount of data transfer when viewed with a plurality of pictures to be encoded.

  Therefore, the present invention has been made in view of such a problem, and provides a motion detection device and a motion detection method that reduce the amount of data transferred to a local memory as seen by a plurality of pictures to be encoded in an image encoding process. For the purpose.

  In order to achieve the above object, a motion detection apparatus according to the present invention detects (a) a motion between pictures in units of a predetermined image area using an encoded reference picture stored in a frame memory. (B) first storage means for storing pixel data of a first encoding target image region belonging to the first encoding target picture, and (c) the first code Second storage means for storing pixel data of a second encoding target image region belonging to a second encoding target picture different from the encoding target picture, and (d) a reference target image region belonging to the reference picture Third storage means for storing pixel data; (e) pixel data of the reference target image region; pixel data of the first encoding target image region; and pixels of the second encoding target image region Each of the data And block matching and further comprising a motion detection means for detecting a motion between pictures respect.

  As a result, the pixel data of the reference target image area is transferred from the frame memory to the third storage means, but motion detection is performed for a plurality of encoding target image areas to which the same reference target image area is referenced. The data transfer amount of the pixel data in the image area can be reduced, and the power consumption can also be reduced.

  Note that the present invention may be realized not only as a motion detection device, but also as a motion detection method for controlling the motion detection device, a motion detection integrated circuit mounted on the motion detection device, and the like. Moreover, it is good also as an inter-screen prediction encoding apparatus provided with a motion detection apparatus.

  As described above, according to the motion detection apparatus of the present invention, the pixel data of the reference target image region is transferred from the external frame memory to the internal local memory for the encoding target image region to be encoded using inter-screen prediction. However, since motion detection is performed for a plurality of encoding target image areas to which the same reference target image area is referenced, the data transfer amount of pixel data in the reference target image area can be reduced, and power consumption can be reduced. Can be reduced.

  In addition, in the generation of prediction errors for a plurality of encoding target image areas, it is not necessary to transfer pixel data of the reference target image area from the external frame memory to the internal local memory, and the transfer amount of the pixel data of the reference target image area is reduced. Can be reduced. Furthermore, in the subsequent processing in the encoding means, it is possible to generate a bit stream that is encoded in the same order as in the conventional encoding technique without affecting the decoding processing.

  Further, according to the motion detection method of the present invention, for the encoding target image region to be encoded using inter-frame prediction, the pixel data of the reference target image region is transferred from the external frame memory to the internal local memory. However, since motion detection is performed for a plurality of encoding target image areas to which the same reference target image area is referenced, the data transfer amount of pixel data in the reference target image area can be reduced, and power consumption can be reduced. Can be reduced.

  Further, according to the motion detection integrated circuit of the present invention, the pixel data of the reference target image area is transferred from the external frame memory to the internal local area for the encoding target image area to be encoded using inter-frame prediction. Although motion detection is performed on a plurality of encoding target image areas that are transferred to the memory but refer to the same reference target image area, the amount of pixel data transferred in the reference target image area can be reduced, and power consumption can be reduced. Can also be reduced.

(Embodiment 1)
Embodiment 1 of the present invention will be described below with reference to the drawings.

  The motion detection apparatus according to the first embodiment is (a) a motion detection apparatus that detects a motion between pictures in units of a predetermined image area using an encoded reference picture stored in a frame memory. (B) a first local memory in which pixel data of a first encoding target image region belonging to the first encoding target picture is stored, and (c) a first encoding target picture different from the first encoding target picture. A second local memory in which pixel data of the second encoding target image region belonging to the second encoding target picture is stored; and (d) a third pixel memory in which the pixel data of the reference target image region belonging to the reference picture is stored. Local memory, (e) the pixel data of the reference target image area, the pixel data of the first encoding target image area, and the pixel data of the second encoding target image area, respectively. Characterized in that it comprises a motion detector for detecting motion between pictures and quenching.

  In addition, the motion detection device according to the present embodiment has the following characteristics (1) to (3).

  (1) The motion detector performs the sequential processing and the parallel processing on the pixel data of the first encoding target image region and the pixel data of the second encoding target image region with respect to the pixel data of the reference target image region. Block match with either.

  (2) The first local memory is characterized in that the first encoding target image area is N (N is a natural number) times a predetermined image area.

  (3) The third local memory stores the reference target image area in units of M (M is a natural number) of pixel data of the first encoding target image area and pixel data of the second encoding target image area. The pixel data is updated.

  Based on the above points, the motion detection apparatus according to the first embodiment will be described. Here, as an example, a series of encoding and decoding functions such as prediction error image generation, frequency conversion, quantization, inverse quantization, inverse frequency, inverse frequency conversion, and restored image are added to the motion detection device. The inter-picture prediction encoding apparatus will be described. In addition, the inter-picture prediction encoding apparatus is compared with the conventional inter-picture prediction encoding apparatus 10 (see FIG. 10), the encoding target block local memory 54 that stores the original image to be encoded in the LSI, and motion detection. The only difference is the unit 13, the motion compensator 14, and the motion vector memory 16, and the other configurations are the same. Therefore, only different configurations will be described, and description of other configurations will be omitted.

  FIG. 1 is a block diagram showing a configuration of an inter-picture prediction encoding apparatus in Embodiment 1 according to the present invention. As shown in the figure, the inter-picture prediction encoding apparatus 100 is compared with the inter-picture prediction encoding apparatus 10 (see FIG. 10), the encoding target block local memory 54, the motion detector 13, and the motion compensator. 14. Instead of the motion vector memory 16, the encoding target block local memories 111 and 112, the motion detector 113, the motion compensator 114, and the motion vector memory 115 are different.

  Here, the external input memory 53 and the external frame memory 51 are external frame memories of the LSI mounting part 110, and the encoding target pictures are stored therein. The encoding target block local memories 111 and 112 and the reference local memory 52 are local memories in the LSI mounting part 50. When the block matching search is performed by the motion detector 113 and when the motion compensation prediction process is performed by the motion compensator 114, Directly accessed. And the part shown with the referential mark 110 among the inter-screen prediction encoding apparatuses 100 is mounted on the LSI.

  In the inter-screen predictive encoding device 100, the motion detector 113 receives the pixel data of the first encoding target image region, the pixel data of the second encoding target image region, and the pixel data of the reference target image region. The motion detection is performed for each of the first encoding target picture and the second encoding target picture.

  When motion detection is performed, the pixel data of the reference target image area is transferred from the external frame memory 51 to the reference local memory 52 via the external bus 61. Further, the pixel data of the first encoding target image area is transferred from the external input memory 53 to the encoding target block local memory 111 via the external bus 62. Pixel data of the second encoding target image area is transferred to the encoding target block local memory 112.

  At this time, as shown in FIG. 2, the reference target image area is a reference target area at a certain time belonging to the picture 120 as shown by the reference area 151, and is the first encoding target image area. This area is commonly referred to when the pixel data and the pixel data of the second encoding target image area are encoded. The first encoding target image area is a 16 × 16 pixel block area to be encoded at a certain point in time belonging to the picture 121 as indicated by the macroblock 141. The second encoding target image area is a 16 × 16 pixel block area to be encoded at a certain point in time belonging to the picture 121 as indicated by the macroblock 142.

  That is, the pixel data of the first encoding target image area belonging to the first encoding target picture is stored in the encoding target block local memory 111. Pixel data of a second encoding target image area belonging to a second encoding target picture different from the first encoding target picture is stored in the encoding target block local memory 112. Reference is made to pixel data of a reference target image area belonging to a reference picture that is commonly referenced when each of pixel data of the first encoding target image area and pixel data of the second encoding target image area is encoded. Stored in the local memory 52.

  Note that the pictures 120 to 123 are illustrated in order of the display time t. The pictures 120 and 123 are encoded pictures, and are referred to when differential encoding of the pictures 121 and 122 is performed. The pictures 121 and 122 are pictures to be coded that have not yet been coded, and the pictures 121 and 122 are also subjected to inter-screen differential coding with the picture 120 as a reference picture (processing 131 and 132).

  Next, the pixel data of the reference target image area is read from the reference local memory 52 via the internal bus 63. Also, pixel data of the first encoding target image area is read from the encoding target block local memory 111 via the internal bus 116. Pixel data of the second encoding target image area is read from the encoding target block local memory 112 via the internal bus 117. Motion detection is performed by the motion detector 113 using these pixel data, and the resulting motion vector information is stored in the motion vector memory 115.

  Further, in the inter-picture prediction encoding apparatus 100, the motion vector obtained by performing motion detection in the motion detector 113 in the motion compensator 114, the first encoding target stored in the encoding block local memory 111. Using pixel data of the image area, pixel data of the second encoding target image area stored in the encoding block local memory 112, and motion compensation reference pixel data which is a part of pixel data of the reference target image area Then, processing for generating motion compensation and a screen prediction error is performed. On the other hand, in the conventional inter-picture prediction encoding apparatus 10, pixel data of an image area referred to when performing motion compensation is also transferred from the external frame memory 51 to the reference local memory 52. That is, the inter-frame prediction encoding apparatus 100 according to the first embodiment performs motion compensation using a part of the pixel data of the reference target image area referred to when motion detection is performed.

  Note that the first encoding target image region is not limited to a unit of macroblocks used in general image encoding processing, and may be a unit including a plurality of macroblocks, for example. The same applies to the second encoding target image region. In addition, if the first encoding target image region and the second encoding target image region are regions in which the reference region of the same encoded picture is referred to when encoding, the same encoding target It may be an area belonging to a picture. Further, when the pixel data of the first encoding target image area is encoded, it is not necessary to refer to the entire reference target image area, and only a part thereof may be referred to. The same applies when the pixel data of the second encoding target image area is encoded. Further, as shown in FIG. 1, instead of making the encoding target block local memories 111 and 112, the reference local memory 52, etc. into individual functional blocks, a plurality or all of them are separated on one memory. It may be realized as a region.

  Further, regarding the encoding process of the intra prediction block in the picture that performs inter prediction such as P picture and B picture, the pixel data of the encoding target image area individually stored in the encoding target block local memories 111 and 112 is obtained. The intra-screen prediction process and a series of subsequent encoding processes may be used in parallel or sequentially. Also, when processing in parallel, the motion detection process that is part of the inter-picture prediction encoding and the intra-picture prediction process that is part of the intra-picture prediction encoding are performed in parallel, and the codes that both processes refer to The target block local memories 111 and 112 may be referred to alternately.

  In addition, the motion detector 113 performs motion detection using pixel data of each of the first encoding target image region, the second encoding target image region, and the reference target image region. If each obtained motion vector is stored in the motion vector memory 115, motion detection for the first encoding target image region and the second encoding target image region may be processed in parallel. After the motion detection for the second encoding target image region is completed, the motion detection for the second encoding target image region may be performed.

  Next, a method for updating the reference local memory 52 and the encoding target block local memories 111 and 112 will be described. Here, a case where the reference local memory 52 and the encoding target block local memories 111 and 112 are updated in units of encoding target macroblocks (16 × 16 pixels) will be described as an example. As shown in FIG. 3A, FIG. 3B, and FIG. 3C, the encoding target image areas (macroblocks) of a plurality of encoding target pictures are sequentially shifted to the right, Along with this, the reference target image area (reference area) of the reference picture has changed.

  First, as shown in FIG. 3A, the first encoding target image region (macroblock 131a) and the second encoding target image region (macroblock 131b) are the same reference target in the same reference picture. Motion detection processing is performed using the image area (reference area 151a) as a reference range. At this time, the pixel data in the reference area 151 a is transferred to the reference local memory 52. Further, the pixel data of the macro block 141 a is transferred to the encoding target block local memory 111, and the pixel data of the macro block 141 b is transferred to the encoding target block local memory 112.

  Next, as shown in FIG. 3B, when the encoding target image area (macroblocks 141a and 142a) moves to the right by one, the reference local memory 52, the encoding target block local memory 111, 112 is updated. At this time, the pixel data of the next reference target image area (reference area 151 b) is transferred to the reference local memory 52. The pixel data of the next first encoding target image area (macroblock 141b) is overwritten in the encoding target block local memory 111, and the pixel data of the next second encoding target image area (macroblock 142b) is encoded. The overwriting target block local memory 112 is overwritten. Further, as shown in FIG. 3C, the reference local memory 52 and the encoding target block local memories 111 and 112 are similarly updated in the next (right adjacent) macroblock.

  Here, according to the conventional inter-picture prediction encoding apparatus 10, when motion detection is performed on the first encoding target image region (macroblock 141a) and the second encoding target image region (macroblock 142a). Then, data transfer occurs for each encoding target image area, and the pixel data of the reference target image area (reference area 151 a) is transferred from the external frame memory 51 to the reference local memory 52. However, according to the inter-frame prediction encoding apparatus 100 in the first embodiment, the first encoding target image region (macroblock 141a) and the second encoding target image are updated by the updating method described with reference to FIG. Since the pixel data of the reference target image area (reference area 151a) only needs to be transferred once from the external frame memory 51 to the reference local memory 52 immediately before performing motion detection on the area (macroblock 142a), data transfer is performed. The amount can be reduced to ½, and power consumption can be reduced.

  An update method different from the update method described with reference to FIG. 3 will be described. As shown in FIGS. 4 (a) and 4 (b), in the reference pictures 161a and 161b, the horizontal positions of the encoding target image areas (macroblocks) are shifted up and down by one macroblock from each other. It is assumed that the reference local memory 52 is updated when the macro block line (horizontal position of the encoding target macro block) is updated.

  As shown in FIG. 4A, in the same way as the updating method described with reference to FIG. 3, for the macroblock lines (macroblocks x1 to x7 at the same horizontal position) belonging to the first encoding target picture. Each time the encoding target image region (macroblock) shifts to the right, the pixel data of the encoding target image region (macroblock) is transferred to the encoding target block local memory 111 each time. Similarly, each time the macroblock changes to the right with respect to the macroblock line (macroblocks y1 to y7 at the same horizontal position) belonging to the second encoding target picture, the pixel data of the macroblock is encoded each time. Transferred to the target block local memory 112. However, the pixel data of the reference area 161a is not transferred until the macro block line is updated, and the reference local memory 52 is not updated.

  Next, as shown in FIG. 4B, the reference target image area in the reference picture moves from the reference area 161a to the reference area 161b at the time when the horizontal position of the encoding target macroblock changes by one macroblock. Therefore, the reference local memory 52 is updated from the pixel data of the reference area 161a to the pixel data of the reference area 161b.

  In the updating method of the reference local memory 52 and the encoding target block local memories 111 and 112 described with reference to FIG. 4, the encoding target block local memories 111 and 112 are updated for each macroblock. The update may be performed in units of a plurality of macro blocks. Further, although the reference local memory 52 is updated when the horizontal position of the macroblock is updated, that is, for each macroblock line, it may be updated for each of a plurality of macroblock lines. For example, H.M. In the H.264 standard, an MBAFF (Macroblock-Adaptive Frame-Field Coding) encoding tool is standardized.

  In MBAFF, encoding processing is performed with two macroblocks as a pair (macroblock pair) vertically. In this case, the pixel data of the same reference area referred to for different macroblock pairs is transferred to the reference local memory 52 at a time, and the reference local memory 52 is updated at the timing when the macroblock pair transitions. It may be done. The reference local memory 52 stores pixel data of a reference area that is referred to an encoding target image area (macroblock) for one macroblock line. However, in this case, a larger capacity of the reference local memory is required as compared with the case described with reference to FIG. Therefore, a temporary cache memory serving as an intermediate buffer is provided in the LSI between the external frame memory 51 and the reference local memory 52, and the entire reference area in the case described with reference to FIG. 4 is temporarily stored in the temporary cache memory. Alternatively, the pixel data of the area referred to the encoding target block may be transferred from the temporary cache to the reference local memory 52.

  If the update method of the reference local memory 52 described with reference to FIG. 4 is used, as in the case described with reference to FIG. The amount of pixel data transferred from the frame memory 51 to the reference local memory 52 can be reduced to ½.

  In addition, although the inter-screen predictive encoding device 100 has been described with two encoding target image regions (the encoding target block local memories 111 and 112), the present invention is not limited to this, and the number of the target encoding regions is not limited to this. Compared to the conventional inter-frame prediction encoding apparatus 10 (see FIG. 10), the pixel data in the reference target image area is prevented from being transferred redundantly, and the reference local area is referenced from the external frame memory 51. A data transfer amount to the memory 52 can be reduced.

  Note that a series of encoding / decoding processes such as prediction error image generation, frequency conversion, quantization, inverse quantization, inverse frequency conversion, and restoration image generation after motion detection are stored in the motion vector memory 115. Using the vector information, encoding is performed in an encoding method and order that can be decoded by a conventional decoding apparatus, that is, similar to the conventional technique. Thereby, even when the inter-picture prediction encoding apparatus 100 according to the first embodiment is used, the same bit stream obtained by the conventional inter-picture prediction encoding apparatus 10 can be obtained.

(Embodiment 2)
Next, a second embodiment according to the present invention will be described with reference to the drawings.

  In the inter-frame predictive coding apparatus according to the second embodiment, in the process of coding the pixel data of the image area referred to when the predicted pixel is generated in the motion compensation process, the coefficient after the quantization, the motion vector difference, and the like are variable. It has a bit information memory for storing information at a certain point of long encoding.

The motion detector may operate in parallel with a plurality of computing resources.
FIG. 5 is a block diagram showing the configuration of the inter-screen predictive coding apparatus according to Embodiment 2 of the present invention. As shown in the figure, inter-frame prediction encoding apparatus 200 has substantially the same configuration as inter-frame prediction encoding apparatus 100 (see FIG. 1) in the first embodiment. Compared with the inter-screen prediction encoding apparatus 100, the difference is that it includes a pre-encoding processor 211, a bit information memory 212, and a post-encoding processor 213. Instead of accumulating motion vectors obtained in motion detection in the motion vector memory 115 for subsequent encoding processing, as in the inter-frame prediction encoding apparatus 100 (see FIG. 1) in the first embodiment. As in the conventional inter-screen predictive coding apparatus 10 (see FIG. 10), the information only for reference as peripheral information is stored in the motion vector memory 16. Other than that, the functions are the same as those indicated by the same reference numerals in FIG.

  In the inter-picture prediction encoding apparatus 200 as well, as in the inter-picture prediction encoding apparatus 100 in the first embodiment, the motion detector 113 uses the pixel data of the first encoding target image area, the second encoding, and so on. Using the pixel data of the target image area and the pixel data of the reference target image area, motion detection is performed for each of the first encoding target picture and the second encoding target picture. Further, in the motion compensator 114, the motion vector obtained by performing motion detection in the motion detector 113, the pixel data of the first encoding target image area stored in the encoding block local memory 111, and the encoding block Using the pixel data of the second encoding target image area stored in the local memory 112 and the reference pixel data for motion compensation that is a part of the pixel data of the reference target image area, motion compensation and screen prediction error are calculated. Generate the process. On the other hand, in the conventional inter-frame prediction encoding apparatus 10, pixel data of an image area referred to when performing motion compensation is transferred from the external frame memory 51 to the reference local memory 52.

That is, in the inter-frame prediction encoding apparatus 200 according to the second embodiment, motion compensation is performed using a part of the pixel data of the reference target image area referred to when motion detection is performed.
Then, similarly to the conventional inter-picture prediction encoding apparatus 10, the motion vector prediction / DifMV calculator 15 performs a process of obtaining a motion vector prediction difference, and further, a frequency converter / quantizer 17, an inverse frequency converter / inverse quantum. The encoder 18 and the regenerator 19 perform each process, and the pre-encoding processor 211 performs pre-processing of the encoding process. In addition, since the pixel regenerated by the regenerator 19 is used as a reference pixel of the subsequent encoding target block, it is stored in the external frame memory 51 as in the conventional technique.

  For example, H.M. In the context-based adaptive binary arithmetic coding (CABAC) of the H.264 coding standard, the pre-encoding processor 211 has a binarization process function and is a post-encoding processor. Reference numeral 213 has an arithmetic encoding process function.

  Next, the bit information after the pre-encoding processor 211 that has performed the series of encoding processes in units of macroblocks is stored in the bit information memory 212. At this time, the encoding target blocks processed in the motion detector 113 are a plurality of macroblocks as shown in the first embodiment, but the processing after motion detection is performed sequentially, and the bit information memory 212 stores Bit information of each macro block is stored in a structure according to the input order to the post-encoding process 213.

  The bit information after the pre-encoding process stored in the bit information memory 212 is output to the outside as the encoded signal 42 similar to the conventional one after the encoding process of one picture to be encoded is completed. Alternatively, in post-encoding processing, H. H.264 is input to the post-encoding processor 213 at the time when bit information sufficient to perform arithmetic encoding processing in the encoding order in accordance with the H.264 standard is stored, and after encoding, 42 is output to the outside.

  Although a series of encoding processes after the motion detection process are sequentially performed here, a series of encoding processors other than the motion detector can be made similar to the conventional one by using this. This can be realized without an increase in. However, when a plurality of motion detections are performed in parallel, a series of subsequent encoding processes may be performed in parallel. In this case, if the number of arithmetic units of each processor is increased, the processing speed can be improved.

  In FIG. 5, the bit information memory 212 is illustrated as an internal local memory of the motion detection device 210. However, like the external frame memory 5151, the bit information memory 212 is configured as an external memory connected to the motion detection device 210. May be.

  According to the second embodiment, it is not necessary to transfer the pixel data of the reference target image area from the external frame memory to the internal local memory in the prediction error generation for the plurality of encoding target image areas, and the reference target image area The amount of transfer of pixel data can be reduced. Furthermore, the motion vector after motion detection is not accumulated, and the bit stream encoded in the same order as in the conventional encoding technique is generated in the subsequent process in the encoding means without affecting the decoding process. It becomes possible.

(Embodiment 3)
Next, a third embodiment according to the present invention will be described with reference to the drawings.

  In the inter-frame prediction encoding apparatus in the third embodiment, a plurality of encoding target pictures refer to a plurality of reference pictures. At this time, when the decoded pixel data generated by performing the decoding process on the pixel data of the first encoding target image area is used as the local decoding pixel data, the motion detector performs the second encoding target image. When performing motion detection on a region, motion detection is performed with reference to local decoded pixel data and decoded pixel data around the local decoded pixel data. In other words, a macro block belonging to a certain coding target picture uses a coded region to which one coding target block belongs as a reference region.

  FIG. 6 is a block diagram showing the configuration of the inter-picture prediction encoding apparatus in Embodiment 3 according to the present invention. As shown in the figure, inter-frame prediction encoding apparatus 300 has substantially the same configuration as inter-frame prediction encoding apparatus 200 (see FIG. 5) in the second embodiment. The only difference is that the regenerated pixel, which is the output from the regenerator 311, is input to the reference local memory 312 compared to the inter-screen predictive coding apparatus 200.

  As a result, as shown in FIG. 7, the reference local memory 312 stores the pixel data of the reference area 351, the reference area 352, and the reference area 353, and the encoding target block local memories 111 and 112 store the pixel data. The pixel data of the macro blocks 343 and 344 are stored. At this time, the regenerator 311 not only outputs the regenerated pixel to the external frame memory 51 but also outputs it to the reference local memory 312. This is because the regenerated image (local decoded image) of the encoding target block 343 is used as a reference area by the encoding target macroblock 345 subsequent to the macroblock 344.

  Note that the pictures 320 to 323 are arranged in the display order and the encoding order. A picture 320 is an encoded picture, and is a picture referenced from the picture 321 and the picture 322. A picture 321 is a picture referenced from the picture 322. As described in the second embodiment, the same reference region or a part of the same reference region is used when block matching is performed on the plurality of encoding target macro blocks by the motion detector 113. For example, an area that is referenced by the macroblock 343 to be encoded for motion detection is the reference area 351 and the reference area 352 of the picture 320, or a part thereof. In addition, an area that is referenced by the macroblock 344 to be encoded in motion detection is a reference area 351 of the picture 320 and a reference area 353 of the picture 321 or a part thereof.

  Note that instead of the picture 321 being a picture that refers to the picture 320, that is, an inter-picture prediction picture (P picture or B picture), for example, the picture 321 may be a picture (I picture) to be subjected to intra-picture prediction encoding. The same configuration can be taken. In this case, the macroblock to be encoded belonging to the picture 322 refers to the encoded area of the picture 321.

  Here, the example in which the regenerated pixels are output from the regenerator 312 to the reference local memory 312 and the external frame memory 51 has been described. However, when it is not necessary to store the regenerated pixel, that is, when it is not referenced from the subsequent macroblock to be encoded, output to the external frame memory 51 is not performed as in the conventional technique. Further, as shown in FIG. 7, when motion detection is performed on two macro blocks to be encoded at the same time, one encoding target block is already included in the picture to which the other encoding target block belongs. When the encoded region is used as a reference region, that is, when the regenerated image of the block to be encoded is referred to as a motion detection target within a certain process after the processing, output to the reference local memory 312 It is not necessary to do. The term “within certain processing” here depends on the capacity of the reference local memory 312, and if this capacity is large, this processing section can be increased. This eliminates the need to output the area to be referred to once to the external frame memory and then transfer it to the reference local memory 312, thereby reducing the amount of data transferred from the outside to the inside of the LSI. Effective in reducing power consumption.

  Further, in the content described with reference to FIG. 7 in the third embodiment, the pixel data of the reference target image area accumulated in the reference local memory 312 is as in the example of FIGS. 3 and 4 of the first embodiment. Areas that are not used as reference areas in subsequent encoding are sequentially overwritten and updated. As a result, the capacity of the internal local memory can be limited to a certain amount.

  As described above, if the example as in the third embodiment is taken, a plurality of blocks to be encoded whose motion is detected using the same reference region or a part of the reference region as a reference region is referred from the external frame memory of the reference region. The amount of data transferred to the local memory (same characteristics as in the second embodiment), and the encoded region of the encoding target picture to which one of the plurality of encoding target blocks belongs is set as the other encoding target block Can be used as a reference area, and the data transfer amount can be further reduced.

(Embodiment 4)
Next, a fourth embodiment according to the present invention will be described with reference to the drawings.

  In the fourth embodiment, an application example of the inter-screen prediction encoding apparatus shown in the first to third embodiments will be described. FIG. 2 is a block diagram of an AV processing unit that realizes an H.264 recorder. FIG. As shown in the figure, an AV processing unit 400 is an AV processing unit such as a DVD recorder or a hard disk recorder that reproduces digitally compressed audio and images.

  The stream data 401 represents audio and image stream data, the image signal 402 represents image data, and the audio signal 403 represents audio data. The bus 410 transfers stream data, audio / image decoded data, and the like. A stream input / output unit 411 is connected to the bus 410 and inputs / outputs stream data 401. The image encoding / decoding unit 412 is connected to the bus 410 and performs encoding and decoding of images.

  The speech coding / decoding unit 413 is connected to the bus 410 and performs speech coding and decoding. The memory 414 stores stream data, encoded data, decoded data, and the like, and includes an area of the external frame memory 51 (see FIG. 1). The memory input / output unit 415 is connected to the bus 410 and is an input / output interface for data signals of the memory 414.

  The image processing unit 416 is connected to the bus 410 and performs pre-processing and post-processing on the image signal. The image input / output unit 417 outputs an image data signal processed by the image processing unit 416 or an image data signal that is only passed without being processed by the image processing unit 416 to the outside as an image signal 402. Further, the image signal 402 from the outside is taken in.

  The audio processing unit 418 is connected to the bus 410 and performs pre-processing and post-processing on the audio signal. The audio input / output unit 419 outputs the audio data signal processed by the audio processing unit 418 or the audio data signal that is only passed without being processed by the audio processing unit 418 to the outside as the audio signal 403. Also, an external audio signal 403 is captured. The AV control unit 420 controls the overall functions of the AV processing unit 400.

  Here, only the encoding operation will be described with reference to FIG. First, the image signal 402 is input to the image input / output unit 417, and the audio signal 403 is input to the audio input / output unit 419.

  Along with this, the AV control unit 420 controls the image processing unit 416 and performs processing such as filter processing and feature amount extraction for encoding using the image signal 402 input to the image input / output unit 417. The data obtained by performing the processing is stored in the memory 414 via the memory input / output unit 415 as original image data. Next, the image encoding / decoding unit 412 is controlled to cause the image encoding / decoding unit 412 to transfer the original image data and the reference image data from the memory 414 via the memory input / output unit 415, and conversely, The image stream data and the local restoration data encoded by the image encoding / decoding unit 412 are transferred from the encoding / decoding unit 412 to the memory 414.

  The image encoding / decoding unit 412 corresponds to almost the entire inter-picture prediction encoding apparatus 100 (see FIG. 1), the image stream corresponds to the signal 42, and the memory 414 includes an external frame memory. 51 areas are included.

  On the other hand, the AV control unit 420 controls the audio processing unit 418 to perform processing such as filtering and feature amount extraction for encoding using the audio signal 403 input to the audio input / output unit 419. The data obtained through the processing is stored in the memory 414 via the memory input / output unit 415 as original audio data. Next, the original audio data is again extracted from the memory 414 via the memory input / output unit 415, encoded, and stored again in the memory 414 as audio stream data.

  Then, the AV control unit 420 processes the image stream data, the audio stream data, and other stream information as one stream data, and outputs the stream data 401 via the stream input / output unit 411. Write to the storage medium.

  Although the operation for one picture has been described in the embodiment, a plurality of inter-screen predictive coding apparatuses are installed in an LSI by dividing one picture into a plurality of areas and processing the divided areas as one picture. You may implement | achieve by the method of mounting and performing parallel processing, or mounting multiple LSIs and performing parallel processing.

  Note that each functional block in the block diagrams (FIG. 1 and FIG. 5) is typically realized as an LSI which is an integrated circuit. These may be individually made into one chip, or may be made into one chip so as to include a part or all of them. For example, the functional blocks other than the memory may be integrated into one chip. However, since the external frame memory 51 (see FIG. 1) and the memory 414 (see FIG. 8) need to store a large amount of data, it is generally mounted with a large capacity DRAM externally attached to the LSI. Is. However, it may be made into one package or one chip due to the improvement of technology.

  The name used here is LSI, but it may also be called IC, system LSI, super LSI, or ultra LSI depending on the degree of integration. Further, the method of circuit integration is not limited to LSI's, and implementation using dedicated circuitry or general purpose processors is also possible. An FPGA (Field Programmable Gate Array) that can be programmed after manufacturing the LSI, or a reconfigurable processor that can reconfigure the connection and setting of circuit cells inside the LSI may be used. Further, if integrated circuit technology comes out to replace LSI's as a result of the advancement of semiconductor technology or a derivative other technology, it is naturally also possible to carry out function block integration using this technology. Biotechnology can be applied.

  The inter-screen predictive coding apparatus according to the present invention can significantly reduce the transfer amount of reference pixels used for motion compensation prediction while suppressing the increase in mounting of the reference local memory. This is effective for realizing a DVD recorder, a hard disk recorder, a camcorder, etc. that handle a large image size using the H.264 standard.

Block diagram of a motion detection apparatus that implements the present invention Schematic diagram showing the relationship between coding blocks and reference pictures in motion detection Schematic diagram showing search area and update area in local memory for reference Schematic diagram showing search area and update area in local memory for reference FIG. 9 is a block diagram showing a configuration of an inter-screen predictive code applicator according to Embodiment 2. FIG. 9 is a block diagram showing a configuration of an inter-screen prediction code applicator according to Embodiment 3. Schematic diagram showing the relationship between coding block and reference area in motion detection H. A block diagram of an AV processing unit realizing a H.264 recorder The block diagram which shows the structure of the conventional inter-screen prediction encoding process The block diagram which shows the structure of the conventional inter-screen prediction decoding apparatus. Schematic diagram showing the configuration of the reference local memory Schematic diagram showing reference pixels transferred per screen

Explanation of symbols

10, 100, 200, 300 Inter-screen prediction encoding device 11 Input memory 12 Multi-frame memory 13 Motion detection unit 14 Motion compensation unit 15 Motion vector prediction / DifMV calculation unit 16, 115 Motion vector memory 17 Frequency conversion / quantization unit 18 Inverse frequency transform / inverse quantization unit 19 Regeneration unit 20 Coding processing unit 50, 110, 210, 310 LSI mounting part 51 External frame memory 52 Reference local memory 53 External input memory 54, 111, 112 Encoding target block local Memory 61, 62 External bus 63, 64 Internal bus 113 Motion detector 114 Motion compensator 116, 117 Internal bus 211 Pre-encoding processor 212 Bit information memory 213 Encoding post-processor 311 Regenerator 312 Reference local memory

Claims (10)

A motion detection device that detects a motion between pictures in units of a predetermined image area using an encoded reference picture stored in a frame memory,
First storage means for storing pixel data of a first encoding target image region belonging to the first encoding target picture;
Second storage means for storing pixel data of a second encoding target image region belonging to a second encoding target picture different from the first encoding target picture;
Third storage means for storing pixel data of a reference target image region belonging to the reference picture;
Block matching is performed on pixel data in the reference target image area, pixel data in the first encoding target image area, and pixel data in the second encoding target image area to detect motion between pictures. And a motion detecting means. The motion detection means includes
Block matching of the pixel data of the first encoding target image region and the pixel data of the second encoding target image region with either sequential processing or parallel processing with respect to the pixel data of the reference target image region The motion detection device according to claim 1, wherein: The first storage means is
The motion detection apparatus according to claim 1, wherein the first encoding target image area is N (N is a natural number) times the predetermined image area. The third storage means is
The pixel data of the reference target image region is updated in units of M (M is a natural number) of pixel data of the first encoding target image region and pixel data of the second encoding target image region. The motion detection device according to claim 1. The motion detection apparatus according to claim 1, wherein the motion detection unit operates in parallel with a plurality of computing resources. In the case where the decoded pixel data generated by performing the decoding process on the pixel data of the first encoding target image region is the locally decoded pixel data,
The motion detection means includes
The motion detection is performed with reference to the local decoded pixel data and decoded pixel data around the local decoded pixel data when performing motion detection on the second encoding target image region. 2. The motion detection device according to 1. The motion detection device according to any one of claims 1 to 6,
An inter-picture prediction encoding apparatus, comprising: encoding means for encoding pixel data of a moving picture based on pixel data decoded using a result obtained by detection by the motion detection apparatus; . The inter-screen predictive encoding device further includes:
A bit information memory for storing bit information after the first half of encoding by the encoding means in units of the predetermined image area;
The inter-picture prediction encoding apparatus according to claim 7, wherein the encoding means performs encoding using the bit information in the second half process. A motion detection method for detecting a motion between pictures in units of a predetermined image area using an encoded reference picture stored in a frame memory,
A first accumulation step of accumulating pixel data of a first encoding target image region belonging to the first encoding target picture in a first storage unit;
A second accumulation step of accumulating pixel data of a second encoding target image area belonging to a second encoding target picture different from the first encoding target picture in a second storage unit;
A third accumulation step of accumulating pixel data of a reference target image area belonging to the reference picture in a third storage unit;
Block matching is performed on pixel data in the reference target image area, pixel data in the first encoding target image area, and pixel data in the second encoding target image area to detect motion between pictures. A motion detection method comprising: a motion detection step. A motion detection integrated circuit that detects motion between pictures in units of a predetermined image area using an encoded reference picture stored in a frame memory,
First storage means for storing pixel data of a first encoding target image region belonging to the first encoding target picture;
Second storage means for storing pixel data of a second encoding target image region belonging to a second encoding target picture different from the first encoding target picture;
Third storage means for storing pixel data of a reference target image region belonging to the reference picture;
Block matching is performed on pixel data in the reference target image area, pixel data in the first encoding target image area, and pixel data in the second encoding target image area to detect motion between pictures. A motion detection integrated circuit.

Images