JP4561688B2 - Moving picture encoding apparatus, moving picture encoding program, moving picture decoding apparatus, and moving picture decoding program - Google Patents

Description

  The present invention relates to a moving image encoding device, a moving image encoding program, a moving image decoding device, and a moving image decoding program. In particular, when encoding a moving image, a boundary that satisfies a condition that satisfies the Poisson equation is defined. A moving picture coding apparatus, a moving picture coding program, a moving picture decoding apparatus, and a moving picture decoding program that are derived based on conditions and a source model suitable for a moving picture and apply the result to coding and decoding About.

  In recent years, various moving image distribution services via digital broadcasting and networks have become common, and in order to record more high-quality and high-quality moving images, further improvement in the coding efficiency of moving images is desired. ing.

  Generally, video information to be transmitted is often multimedia information created by combining digitized voice, image, and other data. The information amount of such video information is generally 1 to 2 bytes when the data is simply digitized and converted. In the case of voice, the telephone quality is 64 Kbits per second. In the case of a moving image, an amount of information of 100 Mbits or more per second is required even for SD (Standard Definition) analog terrestrial television broadcast quality. It is not practical to handle such a large amount of digital information as it is in view of the cost of using the current transmission path and recording medium.

  Therefore, an encoding technique for compressing such an enormous amount of information is currently being actively researched and developed. MPEG (Moving Picture Experts Group) is a representative example of video encoding technology. MPEG is an international standard for moving image data encoding technology. At present, the MPEG standard mainly includes MPEG-1 for storage media of about 1.5 Mbps such as a video CD, and MPEG- which can correspond to various applications such as communication and broadcasting such as digital broadcasting. 2. MPEG-4 that can realize more advanced multimedia contents by object coding and control thereof, and MPEG-7 for content description are defined. By applying such an encoding technique, it is possible to efficiently encode video data having an enormous amount of information and compress the information amount to about 1/20 to 1/40.

  Furthermore, there are some that have introduced a new encoding method such as MPEG-4 AVC (hereinafter referred to as AVC or H.264) which has greatly improved the encoding efficiency of the conventional encoding technology. Thus, more effective information encoding can be performed (see, for example, Non-Patent Document 1 and Non-Patent Document 2). In the encoding method as described above, information compression of video is usually performed by combining MC (motion compensation) and DCT (discrete cosine transform).

  However, in the conventional encoding method, MC is based on the assumption that all pixels in a rectangular area (usually a predetermined square block, hereinafter referred to as MC block or simply block) have the same motion. Since motion vector information is generated by detecting the amount of motion, an image signal may be in a discontinuous state at the MC block boundary in a predicted picture configured by spatially arranging MC blocks. This discontinuous state of the MC block boundary results from the fact that the encoding processing unit such as the search for the motion amount, the subsequent orthogonal transform, and the quantization is the block unit. This phenomenon appears more prominently as the amount of code that can be used at the time of encoding is smaller, and can generally be perceived as block distortion.

  As a method for improving the above-described discontinuous state of the MC block boundary, a process of adaptively smoothing a discontinuous waveform generated between MC blocks for a predicted picture generated by block-based motion compensation Has been proposed (see, for example, Patent Document 1), which can improve the coding efficiency at a low bit rate using motion compensation.

  Next, the configuration and operation of the encoding device and the decoding device when the smoothing process used in Patent Document 1 is introduced in addition to the basic configuration of AVC on behalf of the conventional MPEG encoding device and decoding device. This will be briefly described (see Non-Patent Document 1, Non-Patent Document 2, and Patent Document 1).

  FIG. 22 shows a block diagram of an example of an encoding apparatus when the smoothing process used in Patent Document 1 is introduced to the predicted picture after motion compensation in the AVC encoding apparatus. The basic operation of this encoding apparatus will be described with reference to the flowchart of FIG. In FIG. 22, after an encoded picture as an input image is prepared (step S701), encoded pictures (input code 101) sequentially input are intra-coded by the encoding controller 117 or not. (Step S702), if intra coding is performed, the intra predictor 105 performs intra prediction to generate a predicted picture (Step S703), and the predicted picture is selected by the switch 501. On the other hand, if it is not intra coding, after motion estimation is performed by a motion estimator 502 described later (step S704), motion compensation is performed by a motion compensator 503 described later (step S705), and further smoothed by a smoothing filter 131 described later. (Step S706), the prediction picture obtained thereby is selected by the switch 501.

  The predicted picture selected by the switch 501 is supplied to the differentiator 106, where the redundant component included in the time direction is reduced by taking the difference from the input code 101 which is the coded picture (step S707). ).

  Here, there are three prediction modes for AVC: in-plane prediction (Intra) coding, forward prediction (Predictively) coding, and bi-prediction (Bidirectionaly) coding. In the intra prediction encoding, after using the information in the plane and the prediction information from the plane by the intra predictor 105 without using the output of the motion compensator 503, the conversion by the orthogonal transformer 107 described later is performed. What is coded in this in-plane predictive coding mode is called an I picture. An I picture can be decoded without depending on other pictures at the time of decoding. In AVC, normally, one picture can include a unit of a plurality of slices. However, for the sake of simplicity, it is assumed that one picture is composed of only one slice.

  In forward predictive coding, the motion compensator 503 compensates for a previously coded picture and performs prediction for the current coding target picture. Orthogonal transformation by the orthogonal transformer 107 is performed on the difference picture generated by the difference between the prediction picture generated by this prediction and the encoded picture from the input code 101. As the orthogonal transform performed by the orthogonal transformer 107, DCT (Discrete Cosine Transform) is used in MPEG-1, -2, and 4. Orthogonal transformation can also be performed using other transformation bases such as Hadamard bases and wavelet bases. In AVC, orthogonal transform is performed using integer precision DCT and Hadamard transform. What is encoded in the forward prediction mode is called a P picture. P pictures depend on other pictures and cannot be decoded independently like I pictures, but can have a higher compression rate than I pictures.

  In bi-predictive coding, the motion compensator 503 compensates not only for past and future bi-directional, but also for any two encoded frames, and performs prediction for the current coding target picture. The difference picture generated by performing the difference between the predicted picture generated by this prediction and the encoded picture from the input code 101 is converted by the orthogonal transformer 107. A picture encoded in the bi-prediction mode is called a B picture. The B picture depends on other pictures in the past and the future, so it cannot be decoded independently like the I picture and P picture, but the compression rate can be further increased than the I picture and P picture.

  Each frame and each slice are subjected to prediction processing for each MB (macroblock) of 16 pixels in the horizontal direction and 16 pixels in the vertical direction as a predetermined rectangular area. In addition, prediction processing may be performed by dividing into smaller block units or sub-block units. Hereinafter, in order to simplify the story, the prediction process proceeds on the basis of blocks of 8 pixels in the horizontal direction and 8 pixels in the vertical direction.

  The prediction direction differs depending on the I picture, P picture, and B picture. It is the I picture that encodes all MBs independently. The P picture has two modes: a mode in which encoding is performed by prediction from a past picture, and a mode in which the MB is independently encoded without prediction. In addition, there are four modes in which the MB is independently encoded without prediction, prediction from the future, prediction from the past, prediction from both, and prediction.

  In motion estimation, 1/2 or 1/4 pixel is used to identify an area in which the error is minimized for each block in the encoded picture to be encoded by the motion estimator 502 from the reference picture. A predetermined pattern matching is performed with the accuracy of, and motion information is generated. Here, the motion information is composed of at least motion vector information for identifying a spatial position up to an area corresponding to each block detected by pattern matching and reference picture information for identifying a reference picture. . The motion compensation is performed by the motion compensator 503 generating a predicted picture from the reference picture based on the motion information estimated by the motion estimator 502.

  Orthogonal transformation is performed on the differential picture output from the subtractor 106 by the orthogonal transformer 107 (step S708). The orthogonal transform uses DCT in MPEG-1, -2 and 4, but uses integer precision DCT and Hadamard transform in AVC. DCT is an orthogonal transform that discretely transforms an integral transform based on a cosine function into a finite space. In MPEG-1, -2, and 4, two-dimensional DCT is performed on a DCT block of 8 pixels in the horizontal direction and 8 pixels in the vertical direction, which is a predetermined rectangular area by dividing MB into four. In AVC, two-dimensional integer precision DCT and Hadamard transform are performed on a block of 4 pixels in the horizontal direction and 4 pixels in the vertical direction. In general, since a video signal has many low frequency components and few high frequency components, the coefficients can be concentrated on the low frequency components by performing DCT, and the subsequent quantizer 108 can efficiently reduce the amount of information.

  The DCT coefficient obtained by the orthogonal transformer 107 is quantized by the quantizer 108 (step S709). Quantization is a value obtained by weighting the two-dimensional frequency of 8 pixels in the horizontal direction and 8 pixels in the vertical direction by visual characteristics in MPEG-1, -2 and 4 which are quantization matrices, and 2 in the horizontal direction and 4 pixels in the AVC. The DCT coefficient is divided by the quantized value, with a value obtained by multiplying the dimension frequency weighted by visual characteristics and a value obtained by multiplying the whole of the dimension frequency by a scalar scale value as a quantized value. When inverse quantization is performed by the decoder, a value close to the original DCT coefficient is obtained by multiplying by the quantized value.

  The quantized data output from the quantizer 108 is inversely quantized by the inverse quantizer 109 (step S710). Further, the inverse orthogonal transformer 110 performs inverse orthogonal transform using the transform base corresponding to the orthogonal transformer 107 (step S711). The signal output from the inverse orthogonal transformer 110 is combined with the predicted picture selected from the smoothing filter 131 or the intra predictor 105 selected by the switch 103 in the first combiner 111 (step S712), and then intra. While being supplied to the predictor 105, the deblocking filter 112 performs deblocking processing for reducing block distortion mainly generated by orthogonal transformation or quantization (step S713), and then prepares for reference picture preparation. Therefore, it is stored in the frame memory 104 (step S714). The picture stored in the frame memory 104 is used as a reference picture for obtaining a predicted picture in the motion estimator 502 and the motion compensator 503.

  Here, in the encoding device, in Patent Document 1, in order to improve the discontinuous state of the block boundary in the predicted picture generated by the motion compensator 503 described above, the smoothing filter 131 is placed after the motion compensator 503. By performing a process that adaptively smoothes the discontinuous waveform generated between the blocks for the predicted picture generated by block-unit motion compensation, the discontinuity between blocks is To relieve the situation.

  The smoothed predicted picture subjected to such processing is supplied to the differencer 106 via the switch 501 and also supplied to the first combiner 111 via the switch 103. To reduce the decrease in coding efficiency even when the accuracy of motion estimation in the motion estimator 502 is not sufficient in the conventional AVC or when a block that sufficiently matches in the reference picture cannot be detected. Can do.

  The data quantized by the quantizer 108 is encoded by the entropy encoder 113 (step S715). In general, entropy coding is a VLC (Variable Length Code) device, and performs variable length coding on quantized data. In variable-length coding, in general, zigzag scanning is performed from the low range to the high range of the DCT coefficient, the run length of zero and the effective coefficient value are regarded as one event, and a code with a short code length is assigned from the one with the highest appearance probability. Huffman encoding is performed. The entropy encoder 113 can compress higher information by using arithmetic coding instead of Huffman coding.

  In AVC, CAVLC (Context-Adaptive Variable Length Coding) and CABAC (Context-Adaptive Binary Arithmetic Coding) are used for this entropy coding. Increases efficiency. The entropy-encoded data is multiplexed by the multiplexer 114 (step S716), and output as output code (encoded data) 116 at a predetermined transfer rate from the output transmitter 115 (step S717).

  In FIG. 22, a code amount controller for accumulating encoded data and controlling the code amount can also be provided. The code amount for each macroblock of the encoded data to be output is notified to a code amount controller not shown in FIG. 22, and the error code amount between the notified code amount and the target code amount is quantized. 108 is notified. The quantizer 108 can control the code amount by adjusting the quantization scale based on the notified error code amount information.

  The output code 116 output in this way is distributed as an encoded bit stream by transmission through a storage medium or a network, and is reproduced by a terminal device. In the case of reproduction by the terminal device, the distributed coded bit stream is reproduced after being decoded by the decoding device of the terminal device.

  Next, the decoding apparatus will be described. FIG. 24 shows an example of a decoding apparatus in a case where the smoothing process used in Patent Document 1 is introduced to the predicted picture after motion compensation in the configuration of the AVC decoding apparatus as an example of a conventional decoding apparatus. A block diagram is shown. The basic operation of this decoding apparatus will be described below with reference to the flowchart of FIG.

  In FIG. 24, after the encoded bit stream is received by the input receiver 303 as the input code 301 (step S801), each information such as motion vector information and texture information entropy-encoded by the demultiplexer 304 is obtained. (Step S802), and each information is notified to the entropy decoder 305. The entropy decoder 305 decodes the entropy-encoded texture information (step S803). The decoding controller 313 determines whether or not to perform intra decoding based on the decoding result of the entropy decoder 305 (step S804). If the picture is to be intra decoded, the intra predictor 308 performs intra prediction ( In step S805), the prediction picture obtained thereby is selected by the switch 302. When it is not intra decoding, motion compensation is performed by a motion compensator 317 described later (step S806), and smoothing processing is further performed by the smoothing filter 321 (step S806). In step S807), the prediction picture obtained thereby is selected by the switch 302.

  That is, the entropy decoder 305 notifies the inverse quantizer 306 of the decoding result, while decoding the entropy-encoded motion vector information and notifies the motion compensator 317. In general, entropy decoding is an inverse VLC unit that performs variable length decoding on demultiplexed data. In AVC, modulation modulation is performed by inverse CAVLC or inverse CABAC. Also, the motion compensator 317 generates a predicted picture based on the motion vector information obtained from the entropy decoder 305. The predicted picture is supplied to the smoothing filter 321 and smoothed.

  The inverse quantizer 306 multiplies the entropy-decoded texture information by the quantized value to calculate a value close to the original orthogonal transform coefficient and then supplies the calculated value to the inverse orthogonal transformer 307 (step S808). ). The inverse orthogonal transformer 307 obtains a decoded differential picture by performing inverse orthogonal transform on the supplied orthogonal transform coefficient (step S809). This inverse orthogonal transform uses IDCT (Inverse Discrete Cosine Transform) in MPEG-1, -2, and 4. AVC uses integer precision IDCT and Hadamard transform.

  The difference picture obtained by the inverse orthogonal transformer 307 is supplied to the first synthesizer 309, where it is synthesized with the predicted picture from the smoothing filter 321 or the intra predictor 308 selected by the switch 302 ( Step S810) is supplied to the intra predictor 308, and after deblocking processing for reducing block distortion is performed by the deblocking filter 310 (Step S811), it is stored as a reference picture in the frame memory 311 (Step S810). S812). Also, the decoded image frame is output as the output code 312 from the frame memory 311. In this way, the input code 301 is decoded by the decoding device of the terminal device, and the video information is reproduced.

  Here, in Patent Document 1, when an encoded bitstream is generated, decoding is performed on the encoded bitstream generated by an encoding method that includes a configuration in which smoothing processing is performed on a predicted picture after motion compensation. When performing the above, a smoothing filter 321 that performs the same processing as the smoothing filter processing used in the encoding method is provided in the subsequent stage of the motion compensator 317, so that correct decoding processing can be performed.

  In addition, in recent years, JPEG (Joint Photographic Experts Group) encoding, which is known as a natural image encoding scheme, is further improved in encoding efficiency by using multiple harmonic local orthogonal transform, particularly multiple harmonic local cosine transform. There has been proposed a method for trying (see, for example, Non-Patent Document 3). In this multi-harmonic local cosine transform method, the concept of the Poisson equation is introduced to obtain an estimated signal for the original signal in the block, and a residual signal that is a difference between the original signal and the obtained estimated signal is generated. It is possible to quickly converge DCT coefficients having a higher frequency when introducing, and as a result, it is possible to improve the encoding efficiency.

  FIG. 26 is a conceptual diagram for illustrating the basic concept of this multiple harmonic local cosine transform method. FIG. 26 shows a particular block in the original image frame in which attention is paid to a particular line, and the state of the original signal on the noticed line and the boundary of the original signal on the block boundary is shown in FIG. Shows how it changes.

  With respect to the original signal 2201 shown in FIG. 26A, the inclinations 2202 and 2203 of the image signals at both ends of the original signal corresponding to the block boundaries 2204 and 2205 are obtained. The gradients 2202 and 2203 of the image signals at both ends are used as boundary conditions. In the multiple harmonic local cosine transform method, a predetermined source model that defines the behavior of a signal in a block is applied, and an estimated signal is generated from the signal in the block based on the boundary condition of the block.

  Here, the estimated signal 2206 is generated using the gradients of the image signals at both ends of the original signal as shown in FIG. As the predetermined source model, it is desirable to adopt a model that can estimate a signal closest to the original signal under boundary conditions. In general, in the case of the one-dimensional signal shown in FIG. 26, it can be easily obtained by applying a quadratic function as a predetermined source model, but it is not limited to this model here. Note that. By introducing a predetermined source model in this way, an estimated signal in the block is generated analytically without spending a huge amount of computation and solving the Poisson equation numerically.

  After generating an estimated signal 2206 based on the boundary condition as shown in FIG. 26B, a difference between the original signal and the estimated signal is obtained to obtain a residual signal 2207 as shown in FIG. Is generated. The residual signal 2207 thus obtained is subjected to normal orthogonal transformation, such as DCT, quantization, and entropy coding, thereby completing the natural image coding using the multiple harmonic local cosine transformation method. .

  FIG. 27 and FIG. 28 are intended to show what difference from the normal DCT is caused by using the multiple harmonic local cosine transform method. FIG. 27 shows an example in which normal DCT is performed on the original signal of FIG. In FIG. 27, the horizontal axis indicates x in the x direction of the image signal when attention is paid only to the x direction, and the vertical axis indicates the value (luminance value) of each pixel of one line of the image signal. In FIG. 27, the waveform is shown as a continuous function for convenience, but it is actually a discrete value (the same applies to FIG. 28).

  Since the DCT type normally uses DCT-II, when DCT is performed on a finite original signal divided by blocks as shown in FIG. 27A, the block boundary of the original signal is On one side, here, the original signal is evenly connected with the right side as the axis, and a periodic waveform signal of the original signal as shown in FIG.

  By using such a periodic waveform signal, it can be expressed by a DCT series as shown in FIG. Usually, as indicated by the dotted line portions 2301 to 2303 in FIG. 27B, when the original signal is evenly connected, a portion where the smoothness of the signal is not necessarily sufficient occurs. Due to this influence, the DCT coefficient, which is each amplitude of the DCT series shown in FIG. 27C, is mixed with a signal component different from the signal component originally possessed by the original signal due to the even connection of the signals. Even in a high frequency component region as indicated by the dotted line portion 2304, the DCT coefficient is not necessarily sufficiently converged. Accordingly, when predetermined quantization and entropy coding is performed on such a DCT coefficient, a corresponding code amount is generated.

  On the other hand, FIG. 28 shows the case where the multiple harmonic local cosine transform method is used. After generating an estimated signal 2206 as shown in FIG. 26 (b) using this method, the difference from FIG. 26 (a) is taken to obtain the remaining signal as shown in FIG. 26 (c) and FIG. 28 (a). A difference signal 2207 is obtained. By using this residual signal 2207, as shown in FIG. 28 (a), if this residual signal 2207 is evenly connected with the right side of the signal as an axis, as in FIG. 27 (a), FIG. A periodic waveform signal of the residual signal as shown in FIG.

  The obtained periodic waveform signal of the residual signal is a waveform of a portion corresponding to the periodic waveform signal of the original signal shown in FIG. 27B in the dotted line parts 2401 to 2403 of FIG. As a result, it is possible to improve smoothness and to prevent unnecessary signal components that are not originally included in the original signal from being mixed. Therefore, when the DCT is performed by regarding the periodic waveform signal of the residual signal shown in FIG. 28B as a discrete periodic waveform and the spectrum thereof is expressed, a DCT coefficient sequence as shown by I in FIG. Residual orthogonal transform coefficients (here, residual DCT coefficient information) are obtained. The residual DCT coefficient information indicated by I in FIG. 28C is compared with the DCT coefficient sequence indicated by II in FIG. 28C when ordinary DCT is performed, and a dotted line portion 2404 in FIG. The DCT coefficient is sufficiently suppressed in the high frequency component region as shown, and the state is biased toward the DCT coefficient having a lower frequency component.

  Therefore, by performing predetermined quantization and entropy coding on the DCT coefficients using such a multi-harmonic local cosine transform method, the amount of codes generated can be reduced as compared with the case where coding by normal DCT is performed. This leads to an improvement in coding efficiency.

  The above is the basic concept when the multiple harmonic local cosine transform method is used. Here, in order to briefly explain the basic concept of the multiple harmonic local cosine transform method, an estimated signal is generated based on boundary conditions of the original signal, and a DCT is performed after generating a residual signal based on the difference from the original signal. It was shown that it will be realized.

  Also, the DCT is performed on the original signal in the block and the estimated signal from the boundary condition to obtain the respective DCT coefficients, and then the difference is obtained, as in the case where the DCT is performed on the residual signal. A method for obtaining the result of the above has been conventionally disclosed (see, for example, Non-Patent Document 3).

  Hereinafter, although simple, a method shown in Non-Patent Document 3 will be described. First, an image signal of an image frame is captured as each block shown in FIG.

図２９のΩ^{（０，０）}が現在処理対象となっているブロックであるものとする。また、各ブロック内の各画素は、数２で表現される。

It is assumed that Ω ^{(0, 0) in} FIG. 29 is a block currently being processed. In addition, each pixel in each block is expressed by Equation 2.

まず、各ブロックの画像信号に対して直交変換を行う。ここでは数３に示すようなＤＣＴを行う。

First, orthogonal transformation is performed on the image signal of each block. Here, DCT as shown in Equation 3 is performed.

次に、境界条件である画像信号の傾きを周波数領域で議論することができるように、ＤＣＴ級数で表現する。ここでは数４に示した境界条件のＤＣＴ級数展開の式を利用して画像信号の傾きをｇとした場合に数４に示した式中のＧ_ｋのようなＤＣＴ係数で表現する。

Next, the gradient of the image signal, which is a boundary condition, is expressed by a DCT series so that it can be discussed in the frequency domain. Here, when the gradient of the image signal is set to g using the DCT series expansion formula of the boundary condition shown in Formula 4, it is expressed by a DCT coefficient such as G _{k in} the formula shown in Formula 4.

また、ポアソン方程式の概念を導入してブロック内の原信号に対する推定信号を求める。ここで、ポアソン方程式は、処理対象ブロックＱ_jにおいて、ブロック内の推定信号ｕのラプラシアンである△ｕ_jがソース項Ｋ_jとの間において、次式が成立する方程式である。

In addition, the concept of the Poisson equation is introduced to obtain an estimated signal for the original signal in the block. Here, the Poisson equation, the target block Q _j, between is the Laplacian of the estimated signal u in the block △ u _j is the source term K _j, is an equation in which the following equation is established.

△ u _j = K _j
The estimated signal u in the block can be expressed as shown in Equation 6 using Neumann boundary conditions and DCT series notation. This can be obtained by adding the DCT series expansion components of the estimation signal from each boundary as shown in Equation 5.

このような推定信号ｕに対してＤＣＴを行うことで、次式で表される推定信号Ｕを得ることができる。

By performing DCT on such an estimated signal u, an estimated signal U represented by the following equation can be obtained.

U _{k1, k2} = G _k1 ⁽¹⁾ η _{k1, k2} + G _k1 ⁽²⁾ η ^* _{k1, k2}
+ G _k2 ⁽³⁾ η _{k2, k1} + G _k2 ⁽⁴⁾ η ^* _{k2, k1} (1)
However, in the equation (1), η _{k, m} and η ^* _{k, m} are expressed by the following equations.

ここで、数７に示した式から、η及びη^*はブロック内の位置情報のみに依存し、ブロック内の画像信号には依存しないことから、あらかじめ一意に計算して求めておくことができる。したがって推定信号Ｕを求めるためには、傾きのＤＣＴ係数情報Ｇ_kを求めることが重要となる。

Here, from the equation shown in Equation 7, since η and η ^* depend only on the position information in the block and do not depend on the image signal in the block, they can be uniquely calculated in advance. . Therefore, in order to obtain the estimated signal U, it is important to obtain the DCT coefficient information G _k of the slope.

Here, in order to make it possible to obtain G _k by using DCT coefficient information after performing DCT on the original signal in the block, when obtaining the inclination of the image signal at the block boundary, which is a boundary condition. The slope between the blocks is calculated using the average value of the signal components in each direction in the block as represented by the equation (8) as a representative value, and g is obtained by approximating this slope as the slope at the block boundary.

続いて、この近似されたｇに対してＤＣＴを行うことでＧを求める。ここで、Ｇを求める際に数８に示した式を考慮して数９に示した式変形を行うことで、ＧをＦ、つまりブロック内のＤＣＴ係数情報を利用して求めることが可能となる。

Subsequently, G is obtained by performing DCT on the approximated g. Here, when obtaining G, the equation shown in Equation 9 is considered in consideration of the equation shown in Equation 8, so that G can be obtained using F, that is, DCT coefficient information in the block. Become.

Supervised by Satoshi Okubo, Junya Tsuno, Yoshihiro Kikuchi, Teruhiko Suzuki, “H.264 / AVC Textbook”, Impress Inc., August 2004 Co-authored by Sadayasu Ono, Atshimichi Murakami, and Kotaro Asai, “High-efficiency coding of moving images”, Ohm Corporation, April 2005 Masaru Yamatani, Naoki Saito, “Improvement of DCT-based Compression Algorithms Using Poisson's Equation”, Internet <URL: http://math.ucavis.edu/~saito/publications/phlct.html> Republished patent WO2003 / 003749

  In the conventional video encoding, a search within a predetermined search range corresponding to block matching is performed within a reference image frame when performing motion estimation. In this block matching, normally, a block having the smallest SAD (Sum of Absolute Difference) is searched from the reference image frame, and the block from the current encoding target block is searched. The distance is obtained as motion vector information. Therefore, in the prediction picture generated by performing motion compensation based on the generated motion vector information, the characteristics of the image between the blocks, for example, the shape and contour of the object included in the image, the texture that is the pattern in the block, The property may be in a discontinuous state between blocks without being smoothly connected between blocks.

  Discontinuity of block boundaries and inconsistency of texture information in blocks are generally caused when the motion of an object included in a moving image is large between moving pictures, deformation of the object itself, or multiple objects. It often occurs when there is a change in the interrelationship between the objects, that is, disappearance or appearance of an object.

  In order to improve such discontinuity between blocks and inconsistency of texture information, the invention described in Patent Document 1 described above performs smoothing of an image signal on pixels near the boundary between blocks. Try to alleviate the discontinuous state.

  However, in such an attempt, the discontinuity between blocks may not be sufficiently relaxed due to the strength of smoothing. For example, when the smoothing strength is weaker than the required strength, the discontinuity between blocks cannot be sufficiently improved and the influence remains as block distortion. Also, if the smoothing strength is higher than the required strength, the discontinuity of the block boundary is improved, but the texture information in the block is too smoothed. As a result, the texture information in the block deteriorates.

  In addition, after performing smoothing, predetermined orthogonal transformation, quantization, and entropy coding are performed on the difference image frame generated by performing the difference from the moving image frame to be encoded. . Here, since smoothing is performed on the predicted picture generated by motion compensation after selecting the optimal block in motion estimation, the block is not necessarily the optimal block. There may be others in the picture. Therefore, when trying to obtain more optimal motion vector information in consideration of the amount of code of motion vector information and quantized orthogonal transform coefficient information, motion estimation, motion compensation, and smoothing processing are repeatedly performed to obtain the optimal motion vector information. Vector information needs to be determined, and a large amount of computation must be expended.

  The present invention has been made in view of the above points, and eliminates a discontinuous state between blocks caused by generally used motion estimation and motion compensation without performing a smoothing process. An object of the present invention is to provide a moving picture encoding apparatus, a moving picture encoding program, a moving picture decoding apparatus, and a moving picture decoding program capable of generating a predicted picture in which the continuity of included image signals is maintained. And

  Another object of the present invention is to improve the coding efficiency of the entire coded bitstream by switching the generated difference picture as needed to reduce the amount of information according to the amount of information generated. It is an object to provide a moving image encoding device, a moving image encoding program, a moving image decoding device, and a moving image decoding program.

  In order to achieve the above object, the moving picture coding apparatus according to the first aspect of the present invention subdivides one screen area of the inputted moving picture signal to be coded in units of a rectangular area having a predetermined number of pixels. Generating a difference picture that is a difference signal between a prediction picture generated from a reference picture that is a locally decoded image signal and a coded picture that is a moving image signal to be encoded using a rectangular area as a processing unit, In the moving picture coding apparatus that performs coding, the boundary condition of the boundary portion with the boundary portion between the rectangular region to be predicted in the coded picture and another rectangular region adjacent to the rectangular region as a search unit Is determined by performing a motion vector search in the reference picture with a boundary condition matching the boundary condition, and a rectangle in the encoded picture Region boundary motion estimation means for generating boundary portion motion vector information, which is motion vector information from the boundary portion of the region to the specified boundary portion in the reference picture, and the corresponding boundary portion from the reference picture based on the boundary portion motion vector information The first predicted picture is generated by generating an estimated image signal in each rectangular area in the encoded picture that satisfies the Poisson equation based on the boundary condition of the specified boundary portion. The difference between the region boundary motion compensation unit, the signal component in the rectangular region in the encoded picture to be processed, and the signal component in the rectangular region in the first predicted picture generated by the region boundary motion compensation unit By performing the calculation, reference difference information, which is difference information that becomes a reference for comparison, is generated and becomes a motion vector search target in the reference picture. Difference information between the signal component in the rectangular region and the signal component in the rectangular region in the first predicted picture is generated to generate search target difference information to be compared in motion vector search, A motion vector search is performed within the reference picture for a matching rectangular area while comparing the search target difference information and the signal component of the reference difference information, and a difference component motion vector that is motion vector information up to the rectangular area in the adapted reference picture Based on the difference component motion vector information, the intra-regional difference motion estimation means for generating information, the corresponding rectangular region is identified from the reference picture, the signal component in the rectangular region in the identified reference picture, and the first prediction Intra-regional differential motion compensation means for generating a prediction residual picture that is a differential prediction picture from a difference from a signal component in a corresponding rectangular region in the picture; A first synthesizing unit that generates a second predicted picture by synthesizing the first predicted picture and the predicted residual picture; and a first that generates a first difference picture from a difference between the second predicted picture and the encoded picture. A difference means, an orthogonal transform means for generating orthogonal transform coefficient information by performing a predetermined orthogonal transform on the first difference picture, and a predetermined quantization parameter for the orthogonal transform coefficient information based on a predetermined quantization parameter Quantization means for generating post-quantization information by performing quantization, and performing predetermined dequantization based on a predetermined quantization parameter for the post-quantization information to generate post-quantization information, Decoding difference picture generation means for generating a decoded difference picture by performing predetermined inverse orthogonal transform on the dequantized information, and decoding and local decoding by combining the decoded difference picture and at least the second predicted picture A local decoded image signal generating means for generating an image signal, and holding a local decoded image signal for at least one picture as a reference picture, and holding the reference picture as an area boundary motion estimating means, an area boundary motion compensating means, Predetermined entropy coding is performed on storage means for supplying to each of the intra-regional differential motion estimation means and the intra-regional differential motion compensation means, and at least post-quantization information, boundary motion vector information, and differential component motion vector information And at least an entropy encoding means for generating an encoded bit string and a multiplexing means for generating an encoded bit stream by multiplexing the encoded bit string based on a predetermined syntax structure.

  In the present invention, if a boundary condition in a specific region can be determined, an estimated signal that satisfies the Poisson equation can be generated based on the boundary condition of the region. The boundary condition between the rectangular area in the encoded picture that is the encoding target of the image and the other rectangular area adjacent to the rectangular area is determined as a search unit, and the boundary condition is set as the boundary condition. A boundary part of a reference picture having a matching boundary condition is specified by performing a motion vector search in the reference picture, and a motion vector from the boundary part of the rectangular area in the encoded picture to the specified boundary part in the reference picture Region boundary motion estimation is performed by generating boundary portion motion vector information as information. A boundary signal necessary for generating an estimated signal based on the boundary motion vector information is specified and applied to the Poisson equation to generate a prediction signal, and a first prediction picture is generated to generate a region. Perform boundary motion compensation.

  Furthermore, in the present invention, the reference difference information is generated by calculating the difference between the signal component in the rectangular area in the encoded picture and the signal component in the rectangular area in the first predicted picture, and the reference difference information is generated in the reference picture. Search target difference information to be compared during motion vector search is generated by calculating the difference between the signal component in the rectangular area that is the motion vector search target and the signal component in the rectangular area in the first predicted picture. The search target difference information is compared with the signal component of the reference difference information, and a motion vector search is performed from the reference picture for a suitable rectangular area, and the motion vector information is obtained up to the rectangular area in the adapted reference picture. Intra-region differential motion estimation is performed to generate differential component motion vector information. Then, based on the difference component motion vector information, the corresponding rectangular area is identified from the reference picture, the signal component in the rectangular area in the identified reference picture, and the corresponding rectangular area in the first predicted picture Intra-regional differential motion compensation is performed by generating a prediction residual picture that is a differential prediction picture from the difference from the signal component. A second predicted picture is generated by combining the first predicted picture and the predicted residual picture obtained in this way.

  In order to achieve the above object, the moving picture coding apparatus according to the second invention performs motion vector search by block matching in the reference picture acquired from the storage means in each rectangular area in the coded picture, and is adapted. A motion estimation unit that generates motion vector information up to the rectangular region to be detected, a corresponding rectangular region is identified from the reference picture acquired from the storage unit based on the motion vector information generated by the motion estimation unit, and the fourth prediction Motion compensation means for generating a picture, second difference means for generating a second difference picture from the difference between the fourth predicted picture and the coded picture, a first difference picture, and a second difference picture A difference determination unit that selects a difference picture with a smaller amount of information by comparing based on a predetermined determination criterion, and a difference picture selected by the difference determination unit And an encoding control means for controlling to supply a prediction picture is generated element of the selected difference picture in a predetermined unit, characterized in that added to the first invention.

  In the present invention, after obtaining a fourth predicted picture by performing motion estimation based on a predetermined rectangular area unit to generate motion vector information, and performing motion compensation based on the generated motion vector information A first difference picture is generated from the difference between the encoded picture and the second predicted picture, and a second difference picture is generated from the difference between the encoded picture and the fourth predicted picture. Generated because the difference picture and the second difference picture are compared by the difference judging means based on a predetermined judgment condition, the difference picture with the smaller amount of information is selected and supplied to the subsequent orthogonal transformation The difference picture can be switched at any time so as to reduce the information amount.

  In order to achieve the above object, the moving picture coding program of the third invention subdivides one screen area of the inputted moving picture signal to be coded in units of a rectangular area having a predetermined number of pixels. Dividing and generating a difference picture that is a difference signal between a prediction picture generated from a reference picture that is a locally decoded image signal and a coded picture that is a moving image signal to be encoded, with a rectangular area as a processing unit, In a moving picture coding program for causing a computer to perform coding, the computer is caused to function as each component of the constituent elements of the moving picture coding apparatus of the first invention. . In order to achieve the above object, a moving picture coding program according to a fourth aspect of the invention causes a computer to function as each means of components of the moving picture coding apparatus according to the third aspect of the invention.

  In order to achieve the above object, a moving picture decoding apparatus according to a fifth invention is a moving picture coding apparatus according to the first or second invention, or a moving picture coding program according to the third or fourth invention. The encoded bit stream generated by the computer that executed the above is acquired from a predetermined storage medium or a predetermined transmission path, and the decoded moving image signal is obtained by performing a decoding operation on the acquired encoded bit stream. An output video decoding apparatus, wherein a demultiplexing unit that demultiplexes encoded information based on a predetermined syntax structure with respect to an encoded bitstream, and a demultiplexing acquired from the demultiplexing unit A predetermined entropy decoding is performed on the subsequent information, and at least the post-quantization information, boundary motion vector information, difference component motion vector information, and a predetermined syntax structure are configured. Entropy decoding means for generating parameter information necessary for the generation, and post-quantization information is generated by performing predetermined inverse quantization based on predetermined quantization parameters for the post-quantization information acquired from the entropy decoding means Then, by performing predetermined inverse orthogonal transform on the dequantized information, the decoding difference picture generating means for generating a decoding difference picture, and the decoding difference picture and at least the prediction picture for synthesis are combined and local decoding is performed. First reference means for generating an image signal, storage means for holding a locally decoded image signal for at least one picture as a reference picture, and boundary motion vector information acquired from the entropy decoding means, Identify the boundary condition of the corresponding boundary part, and satisfy the Poisson equation based on the boundary condition of the identified boundary part The region boundary motion compensation unit that generates the first predicted picture by generating the estimated image signal in each rectangular region in the encoded picture that is the moving image signal to be encoded, and the difference acquired from the entropy decoding unit Based on the component motion vector information, a corresponding rectangular region is identified from the reference picture, and a signal component in the rectangular region in the identified reference picture and a signal component in the corresponding rectangular region in the first predicted picture By synthesizing the intra-regional differential motion compensation means for generating a prediction residual picture that is a differential prediction picture from the difference, the first prediction picture and the prediction residual picture, the first synthesis means is used for the above-described prediction for synthesis. And at least second combining means for generating a second predicted picture supplied as a picture.

  In the present invention, a coded bit stream generated by the moving picture coding apparatus according to the first or second invention or the computer executing the moving picture coding program according to the third or fourth invention is transferred to a predetermined transmission line or By obtaining from the recording medium and entropy decoding, at least post-quantization information, boundary motion vector information, difference component motion vector information, and parameter information necessary to configure a predetermined syntax structure are generated, A decoded differential picture is generated by sequentially performing inverse quantization and inverse orthogonal transform on the post-quantization information, and a local decoded image signal is generated by combining the decoded differential picture and at least a prediction picture for synthesis. . The above-mentioned prediction picture for synthesis is encoded such that the boundary condition of the corresponding boundary part is specified from the reference picture based on the boundary motion vector information, and the Poisson equation is satisfied based on the boundary condition of the specified boundary part It is generated by synthesizing the first predicted picture obtained by generating the estimated image signal in each rectangular area in the encoded picture that is the target moving image signal and the prediction residual picture. The prediction residual picture specifies a corresponding rectangular area from the reference picture based on the difference component motion vector information, and the signal component in the rectangular area in the specified reference picture corresponds to the corresponding one in the first prediction picture. It is generated from the difference from the signal component in the rectangular area.

  In order to achieve the above object, in the moving picture decoding apparatus according to the sixth aspect of the invention, the entropy decoding means of the fifth aspect of the invention further has a function of decoding motion vector information, and further entropy decoding means. Necessary for constructing a syntax structure decoded by the motion compensation means and entropy decoding means for identifying the corresponding rectangular area from the reference picture based on the motion vector information obtained from the above and generating the fourth predicted picture Decoding control information necessary for decoding control is obtained from the parameter information, and the second predicted picture and the fourth predicted picture are combined as predicted pictures to be supplied to the first synthesizing unit according to the decoding control information. A decoding control means for selecting one is provided.

  In order to achieve the above object, the moving picture decoding program of the seventh invention is a moving picture coding apparatus of the first or second invention, or a moving picture coding program of the third or fourth invention. The encoded bit stream generated by the computer that executed the above is acquired from a predetermined storage medium or a predetermined transmission path, and the decoded moving image signal is obtained by performing a decoding operation on the acquired encoded bit stream. A moving picture decoding program for causing a computer to execute the output is characterized by causing the computer to function as each component of the constituent elements of the moving picture decoding apparatus of the fifth invention. In order to achieve the above object, a moving picture decoding program according to an eighth aspect of the invention causes a computer to function as each means of components of a moving picture decoding apparatus according to the sixth aspect of the invention.

  According to the first or third aspect of the invention, the region boundary motion compensation is performed by generating the first predicted picture, the intra-regional difference motion estimation for generating the difference component motion vector information is performed, and the difference component motion vector is generated. Based on the information, the corresponding rectangular area is identified from the reference picture, and a difference is obtained from the difference between the signal component in the rectangular area in the identified reference picture and the signal component in the corresponding rectangular area in the first predicted picture. Intra-region differential motion compensation is performed by generating a prediction residual picture that is a prediction picture, and a second prediction picture is generated by synthesizing the first prediction picture and the prediction residual picture obtained thereby. Since the predetermined orthogonal transform is performed on the first difference picture generated from the difference between the second predicted picture and the encoded picture, generally used motion Caused by Jooyobi motion compensation, to eliminate the discontinuous state between blocks, in the block boundary portion, resulting an effect of continuity of the image signal containing it is possible to generate a prediction picture, as is maintained.

  Also, according to the first or third invention, a prediction residual picture is generated by predicting the remaining image signal components that could not be sufficiently predicted by the first prediction picture from the reference picture again, Since the second predicted picture is generated by synthesizing with one predicted picture, there is an effect that a predicted picture closer to the encoded picture can be generated.

  Furthermore, according to the first or third invention, the discontinuous state between the rectangular regions is eliminated, and the second predicted picture closer to the encoded picture including the continuous image signal is obtained. The effect of motion estimation and motion compensation on the quality of the difference picture generated by the difference between the predicted picture and the encoded picture is eliminated, and the efficiency of orthogonal transformation and quantization in which the subsequent difference picture is supplied is reduced. The effect of improving is acquired.

  Furthermore, according to the first or third invention, when performing the region boundary motion estimation, the inclination of the image signal, which is the boundary condition corresponding to each side of the rectangular region, is moved separately for each side of the rectangular region. Even if the image signal included in the picture is deformed between the coded picture and the reference picture by performing vector search and generating each boundary part motion vector information, separate corresponding to each side A first prediction picture with higher quality than the conventional motion compensation is generated from the inclination of the image signal specified from the boundary motion vector information of the image, and a prediction residual picture for further approaching the encoded picture is synthesized Thus, encoding efficiency can be further improved.

  In addition, according to the second or fourth invention, by performing motion vector search by a predetermined rectangular area unit to generate motion vector information, and performing motion compensation based on the generated motion vector information. After obtaining the fourth predicted picture, a first difference picture is generated from the difference between the encoded picture and the second predicted picture, and the second difference is calculated from the difference between the encoded picture and the fourth predicted picture. A picture is generated, the first difference picture and the second difference picture are compared based on a predetermined determination condition by a difference determination means, and a difference picture having a smaller amount of information is selected. Since it is switched from time to time to reduce the amount of information that generates differential pictures, the coding efficiency of the entire coded bitstream can be improved. .

  Further, according to any of the fifth to eighth inventions, an encoded bit stream generated by the moving image encoding apparatus or the moving image encoding program of the present invention is acquired from a predetermined transmission path or a recording medium, Efficiently transmit, receive, and play back an encoded bitstream having a smaller code amount than the prior art by decoding with the computer using the video decoding device of the present invention or the video decoding program of the present invention. It has the feature that can be made possible.

  Next, the best mode for carrying out the invention will be described with reference to the drawings.

(First embodiment)
First, a first embodiment of the video encoding device and video decoding device of the present invention will be described. FIG. 1 shows a block diagram of a first embodiment of a moving picture coding apparatus according to the present invention. In the figure, the same components as those in FIG. 22 are denoted by the same reference numerals. As shown in FIG. 1, the moving picture coding apparatus according to the present embodiment includes a switch 102, a switch 103, a frame memory 104, an orthogonal transformer 107, a quantizer 108, an inverse quantizer 109, and an inverse orthogonal transformer 110. , First synthesizer 111, entropy encoder 113, multiplexer 114, output transmitter 115, encoding controller 117, region boundary motion estimator 118, region boundary motion compensator 119, first difference unit 120, region An inner differential motion estimator 121, an intra-region differential motion compensator 122, a second synthesizer 123, and a second differentiator 124 are provided. The second subtractor 124 performs the same operation as that of the subtractor 106 shown in FIG. Further, as shown in FIG. 1, it is desirable to have a configuration including an intra predictor 105 and a deblocking filter 112.

  Next, details of each component of the first embodiment of the moving picture coding apparatus of the present invention shown in FIG. 1 will be described below. The switch 102 includes a first difference picture (generated from an encoded picture and a second predicted picture) supplied from the first subtractor 120, and a second difference picture (encoding) supplied from the second subtractor 124. A function of supplying a necessary difference picture to the orthogonal transformer 107 by switching between a picture and a third predicted picture by intra prediction) according to an instruction from the encoding controller 117.

  The switch 103 is necessary by switching the second predicted picture supplied from the second synthesizer 123 and the third predicted picture supplied from the intra predictor 105 in accordance with an instruction from the coding controller 117. A function of supplying a predicted picture to the first synthesizer 111.

  The orthogonal transformer 107 has a function of acquiring a difference picture from the switch 102, a function of generating orthogonal transform coefficient information by performing predetermined orthogonal transformation on the acquired difference picture, and supplying the information to the quantizer 108. Have Here, it is desirable that the orthogonal transform base used in the predetermined orthogonal transform is a DCT (discrete cosine transform) base, and the generated orthogonal transform coefficient information is DCT coefficient information. In the following description of the present embodiment, it is assumed that DCT is used for the orthogonal transform base, but it should be noted that the present invention is not particularly limited by this orthogonal transform base.

  The quantizer 108 obtains DCT coefficient information from the orthogonal transformer 107, performs predetermined quantization based on a predetermined quantization parameter, generates post-quantization information, and performs inverse quantization 109 and entropy. It has a function of supplying to the encoder 113. The inverse quantizer 109 obtains the post-quantization information from the quantizer 108 and performs a predetermined inverse quantization on the acquired post-quantization information based on a predetermined quantization parameter. A function of generating post-quantization information and supplying the information to the inverse orthogonal transformer 110.

  The inverse orthogonal transformer 110 generates a decoded differential picture by performing a predetermined inverse orthogonal transform on the acquired inverse quantized information from the inverse quantizer 109 and a function of obtaining the inverse quantized information. A function of supplying to the first combiner 111. The first combiner 111 has a function of acquiring a decoded differential picture from the inverse orthogonal transformer 110 and acquiring a predicted picture from the switch 103. A local decoded picture (hereinafter simply referred to as a decoded picture) is generated by synthesizing the acquired decoded differential picture and the predicted picture. In FIG. It has a function to supply. The deblocking filter 112 acquires a decoded picture from the first combiner 111, performs a predetermined deblocking filter process, and then supplies the decoded picture to the frame memory 104 to store the decoded picture as a reference picture It has the function to do.

  In FIG. 1, the frame memory 104 has a function of acquiring a decoded picture after the deblocking filter processing from the deblocking filter 112 and storing it as a reference picture. Also, a coded picture that is an input code may have a function of storing it as a reference picture according to a coding mode. Further, the frame memory 104 has a function of supplying a requested reference picture to each unit that requires a reference picture. In the frame memory 104 in FIG. 1, reference pictures required in response to a request are converted into an intra predictor 105, a region boundary motion estimator 118, a region boundary motion compensator 119, an intra region differential motion estimator 121, and an intra region differential motion compensation. At least a function of supplying to the device 122.

  The intra predictor 105 has a function of acquiring a decoded picture from the first combiner 111 and generating a third predicted picture by performing predetermined intra prediction. Note that the configuration may be such that a picture used for intra prediction is acquired from the frame memory 104. The intra predictor 105 has a function of supplying the generated third predicted picture to the terminal b of the switch 103 and the second subtractor 124.

  The entropy encoder 113 obtains at least post-quantization information from the quantizer 108, boundary motion vector information from the region boundary motion estimator 118, and difference component motion vector information from the intra-region difference motion estimator 121. It has the function to do. Here, it is desirable that the entropy encoder 113 has a function of acquiring parameter information and the like necessary for configuring other predetermined syntax structures from each unit of the encoding device. Parameter information necessary to configure a predetermined syntax structure includes macroblock information for specifying various states of the macroblock, quantization parameter information used when performing quantization and inverse quantization, and an intra prediction method It is further preferable that intra prediction mode information for specifying the frame order information, frame order information for specifying the reference order of the reference image frames, and the like. The entropy encoder 113 further has a function of generating an encoded bit string by performing predetermined entropy encoding on the acquired various pieces of information and supplying the encoded bit string to the multiplexer 114.

  The multiplexer 114 obtains an encoded bit string from the entropy encoder 113, generates an encoded bit stream by performing predetermined multiplexing based on a predetermined syntax structure, and outputs and transmits the encoded bit stream. A function of supplying to the container 115. The output transmitter 115 acquires an encoded bit stream from the multiplexer 114, and outputs the acquired encoded bit stream as an output code to a predetermined transmission path, a predetermined recording medium, or the like. It has a function of generating packet information by performing packetization processing so as to have a configuration having a payload.

  Here, the packet information treats the packet header and payload as one unit. Usually, the obtained encoded bit stream is divided into smaller units, stored in a payload, and packetized by adding a predetermined packet header. The output transmitter 115 has a function of outputting each generated packet information as an output code according to the state of a transmission path or a recording medium as a predetermined output destination.

  The encoding controller 117 supplies various parameter information necessary for encoding to each unit constituting the encoding device and controls various parameters used in order to control the operation of the encoding device of the present embodiment. It has a function of supplying information to the entropy encoder 113, a function of controlling input / output of each unit, and a function of performing switching control of at least the switch 102 and the switch 103.

  The region boundary motion estimator 118 has a function of acquiring an encoded picture that is an encoding target in a moving image frame input as the input code 101 and a reference picture stored in the frame memory 104. The region boundary motion estimator 118 uses a feature that if a boundary condition in a specific region can be determined, an estimation signal that satisfies the Poisson equation can be generated based on the boundary condition of the region. In addition, the obtained encoded picture is divided into predetermined block areas, and the function of calculating the inclination of the image signal of each side, which is a boundary condition of each side between the blocks in the encoded picture, is provided.

  Further, the region boundary motion estimator 118 calculates the inclination of the image signal corresponding to each side from the position having the inclination of the image signal closest to the inclination of the image signal of each side in the encoded picture, while referring to the reference picture. The region boundary motion vector information is generated by performing the motion vector search in the region, the boundary portion motion vector information is generated, the region boundary motion estimation is performed, and the generated boundary portion motion vector information is supplied to the region boundary motion compensator 119. Furthermore, the region boundary motion estimator 118 has a function of supplying the generated boundary portion motion vector information to the entropy encoder 113.

  The region boundary motion compensator 119 is configured to obtain the estimation signal in the block based on the boundary motion vector information from the region boundary motion estimator 118, the function of acquiring the reference picture from the frame memory 104, and the acquired boundary motion vector information. By specifying the slope of the image signal on each side required from the reference picture and applying the slope of the image signal on each side as a boundary condition to the Poisson equation And generating a first prediction picture that is a prediction signal of each block in the picture, thereby performing region boundary motion compensation. The region boundary motion compensator 119 has a function of supplying the generated first prediction picture to the second synthesizer 123, the intra-regional difference motion estimator 121, and the intra-regional difference motion compensator 122.

  The first subtractor 120 has a coded picture that is a coding target of a moving image that is the input code 101, a function of obtaining a second predicted picture from the second synthesizer 123, and the obtained coded picture The first difference picture is generated based on the difference from the second predicted picture, and the generated first difference picture is supplied to the terminal a of the switch 102.

  The intra-regional difference motion estimator 121 includes a function of obtaining a coded picture that is an encoding target of a moving image that is the input code 101, a reference picture stored in the frame memory 104, and a region boundary motion compensator. 119 to obtain a first predicted picture. In addition, the intra-region difference motion estimator 121 includes a signal component in a rectangular region in the coded picture to be processed, that is, a block in this case, and the first predicted picture generated by the region boundary motion compensator 119. It has a function of generating reference difference information, which is difference information serving as a reference when performing comparison, by performing a difference calculation with a signal component in the block.

  Further, the intra-region difference motion estimator 121 performs a difference operation between a signal component in a block that is a motion vector search target in the reference picture and a signal component in the block in the first predicted picture. , Generating search target difference information to be compared at the time of motion vector search, performing a motion vector search from the reference picture while comparing the search target difference information and the signal component of the reference difference information, and matching It has a function of generating difference component motion vector information which is motion vector information up to a rectangular area in the reference picture.

  Here, the motion vector search for generating the difference component motion vector information is repeatedly performed from time to time within the search range in the reference picture for a suitable region while moving the comparison target region with a predetermined motion accuracy. While identifying. Further, the intra-region difference motion estimator 121 has a function of supplying the generated difference component motion vector information to the intra-region difference motion compensator 122 and the entropy encoder 113.

  The intra-region difference motion compensator 122 has a function of acquiring the corresponding reference picture from the frame memory 104, the first predicted picture from the region boundary motion compensator 119, and the difference component motion vector information from the intra-region difference motion estimator 121. And the corresponding block area from the reference picture based on the acquired difference component motion vector information, the signal component in the block area in the identified reference picture, and the corresponding rectangular area in the first predicted picture And a function of generating a prediction residual picture that is a differential prediction picture from a difference from the signal component of the signal component, and a function of supplying the generated prediction residual picture to the second synthesizer 123.

  The second synthesizer 123 has a function of acquiring the first predicted picture from the regional boundary motion compensator 119 and a prediction residual picture from the intra-regional difference motion compensator 122, and the acquired first predicted picture and prediction residual A function of combining a picture and generating a second predicted picture; and a function of supplying the generated second predicted picture to the terminal a side of the switch 103 and the first subtractor 120.

  The second subtractor 124 calculates the second from the difference between the encoded picture that is the input code 101 and the third predictive picture from the intra predictor 105 and the acquired encoded picture and the third predictive picture. The difference picture is generated and supplied to the terminal b of the switch 102.

  Next, the operation of the first embodiment of the moving picture coding apparatus of the present invention shown in FIG. 1 will be described with reference to the flowchart of FIG. First, as an input code 101, a frame or a field that is an encoding target of a moving image is prepared as an encoded picture (step S101). Subsequently, the encoding controller 117 determines whether the current encoding mode of the encoding device is intra encoding or not intra encoding (step S102).

  In the case of intra coding, the coding controller 117 supplies parameter information for performing intra coding to each unit of the coding apparatus, connects the switch 102 to the terminal b, and switches the switch 103 to By connecting to the terminal b, preparation for performing intra coding is performed. Thereafter, the intra predictor 105 performs predetermined intra prediction (step S103). Thereafter, the process proceeds to step S108.

  On the other hand, when the encoding is not intra encoding, parameter information for performing inter encoding is supplied to each unit, and the switch 102 is connected to the terminal a, and the switch 103 is connected to the terminal a, thereby performing inter encoding. Prepare for. Thereafter, the region boundary motion estimator 118 acquires a coded picture prepared as an input code and a reference picture stored in the frame memory 104, and performs predetermined region boundary motion estimation (step S104). The boundary motion vector information generated thereby is supplied to the region boundary motion compensator 119 and also to the entropy encoder 113.

  The region boundary motion compensator 119 acquires boundary region motion vector information from the region boundary motion estimator 118 and the corresponding reference picture from the frame memory 104, and performs predetermined region boundary motion compensation (step S105). The first predicted picture generated in this way is supplied to the second synthesizer 123, the intra-regional difference motion estimator 121, and the intra-regional difference motion compensator 122.

The intra-region difference motion estimator 121 acquires the encoded picture to be encoded as the input code 101, and also obtains the corresponding reference picture from the frame memory 104 and the first predicted picture from the region boundary motion compensator 119. Obtaining and performing intra-regional differential motion estimation (step S106). Thereby, difference component motion vector information is generated, and the difference component motion vector information is supplied to the intra-region difference motion compensator 122 and the entropy encoder 113.
The intra-regional difference motion compensator 122 acquires the corresponding reference picture from the frame memory 104, the first predicted picture from the region boundary motion compensator 119, and the difference component motion vector information from the intra-regional difference motion estimator 121, respectively. Then, intra-region differential motion compensation is performed (step S107). As a result, a prediction residual picture is generated, and the prediction residual picture is supplied to the second synthesizer 123.

  Thereafter, the second synthesizer 123 synthesizes the first predicted picture from the regional boundary motion compensator 119 and the predicted residual picture from the intra-regional difference motion compensator 122 to generate a second predicted picture. The second predicted picture is supplied to the terminal a side of the switch 103 and the first subtractor 120. Thereafter, the process proceeds to step S108.

  In step S108, a difference calculation is performed between the third predicted picture generated by the intra predictor 105 or the second predicted picture generated by the second synthesizer 123 and the encoded picture that is the input code 101 (step S108). S108). Here, when the encoding mode is intra encoding, the second subtractor 124 acquires the encoded picture and the third predicted picture generated by the intra predictor 105, and the acquired encoded picture and By performing a difference operation with the third predicted picture, a second difference picture is generated and supplied to the orthogonal transformer 107 via the switch 102 connected to the terminal b.

  On the other hand, when the encoding mode is inter encoding, the first subtractor 120 acquires the encoded picture and the second predicted picture generated by the second synthesizer 123, and the acquired encoded picture and A first difference picture is generated by performing a difference calculation with the second predicted picture, and is supplied to the orthogonal transformer 107 via the switch 102 connected to the terminal a.

  The orthogonal transformer 107 acquires the difference picture selected by the switch 102 and performs a predetermined orthogonal transform (step S109). Thereby, orthogonal transform coefficient information (here, DCT coefficient information) is generated, and the DCT coefficient information is supplied from the orthogonal transformer 107 to the quantizer 108.

  The quantizer 108 acquires DCT coefficient information from the orthogonal transformer 107, and performs predetermined quantization based on a predetermined quantization parameter (step S110). Thus, post-quantization information is generated, and the post-quantization information is supplied to the inverse quantizer 109 and the entropy encoder 113, respectively. The inverse quantizer 109 acquires post-quantization information from the quantizer 108 and performs predetermined inverse quantization based on a predetermined quantization parameter (step S111). As a result, information after inverse quantization is generated, and the information after inverse quantization is supplied to the inverse orthogonal transformer 110.

  The inverse orthogonal transformer 110 acquires information after inverse quantization from the inverse quantizer 109, performs predetermined inverse orthogonal transform (step S112), generates a decoded difference picture, and supplies the decoded difference picture to the first combiner 111. The first synthesizer 111 acquires a decoded differential picture from the inverse orthogonal transformer 110 and a predicted picture from the switch 103, combines the acquired decoded differential picture and the predicted picture (step S113), and is a locally decoded image signal. A decoded picture is generated, and the decoded picture is supplied to the deblocking filter 112. When the encoding mode is intra encoding, the decoded picture is further supplied to the intra predictor 105.

  The deblock filter 112 acquires a decoded picture from the first combiner 111, performs a predetermined deblock filter process (step S114), and generates a decoded picture after the deblock filter process. The generated decoded picture after the deblocking filter process is stored in the frame memory 104 as a reference picture in order to prepare a reference picture for the next encoding (step S115).

  Thereafter, in order to output a series of encoding results, the entropy encoder 113 includes at least quantized information from the quantizer 108, boundary motion vector information from the region boundary motion estimator 118, and intra-regional differential motion estimator. The difference component motion vector information is acquired from 121, and predetermined entropy encoding is performed (step S116) to generate an encoded bit string. Here, the entropy encoder 113 may further acquire various parameter information used at the time of encoding from each unit of the encoding device and perform predetermined entropy encoding. The generated encoded bit string is supplied to the multiplexer 114.

  The multiplexer 114 acquires the encoded bit string from the entropy encoder 113, multiplexes based on a predetermined syntax structure (step S117), and generates an encoded bit stream. The generated encoded bit stream is supplied to the output transmitter 115. The output transmitter 115 acquires the multiplexed bit stream from the multiplexing unit 114, performs a predetermined packetization process, etc., and then outputs and transmits the encoding result to a predetermined transmission path or a predetermined recording medium (step S118). ).

  By performing the processes of steps S101 to S118 as described above, the encoding process of the encoded picture that is the encoding target at a certain point in time is completed in the first embodiment of the encoding apparatus of the present invention. To do. As described above, in the present embodiment, an estimation signal is generated from the Poisson equation based on the boundary condition of the block boundary, and further prediction is performed by performing motion estimation on the remaining residual signal, By generating 2 prediction pictures, discontinuity between blocks can be eliminated and encoding efficiency can be improved.

  Next, with respect to the region boundary motion estimation process performed by the region boundary motion estimator 118 included in the moving image encoding apparatus according to the present embodiment and the region boundary motion compensation process included in the region boundary motion compensator 119, FIGS. The details will be described.

  FIG. 13 is a conceptual diagram showing a state of motion estimation in which motion vector information is generated by performing motion vector search by general block matching. Here, for the sake of simplicity, the size of the predetermined rectangular area used for block matching is assumed to be a block of 4 pixels in the horizontal direction and 4 pixels in the vertical direction.

First, the encoded picture 1301 is divided into blocks, and attention is paid to the block of the prediction target 1303 for which a current motion vector search is to be performed. Here, the boundary portion between the block of the prediction target 1303 and the surrounding blocks is represented as Γ ⁽¹⁾ , Γ ⁽²⁾ , Γ ⁽³⁾ , Γ ⁽⁴⁾ . Also, in the reference picture 1302, a region existing at the same spatial position as the block of the prediction target 1303 in the encoded picture 1301 is set as a reference position 1304, and motion vector search is started from this reference position 1304.

  In the motion vector search, normally, a search range 1305 in the reference picture 1302 is set, and block matching is performed within the search range 1305 with 1/2 pixel accuracy or 1/4 pixel accuracy. In block matching, usually, a suitable block 1306 is specified by searching for a block having a minimum SAD (Sum of Absolute Difference) from the search range 1305 of the reference picture 1302. Thereafter, motion estimation is performed by obtaining a difference in spatial position between the block of the prediction target 1303 and the matched block 1306 as motion vector information 1307.

  Accordingly, it is possible to identify a suitable block 1306 in the reference picture 1302 based on the generated motion vector information 1307, duplicate the pixel information in this block 1306, and place it in the corresponding spatial position of the predicted picture. Motion compensation is performed for each block. Since motion compensation is performed in units of blocks in this way, the characteristics of the image between each block, such as the shape and contour of the object included in the image, and the nature of the texture that is the pattern in the block, are smoothly connected between the blocks. Without being discontinuous between blocks.

FIG. 14 is a conceptual diagram showing a state of region boundary motion estimation processing provided in the region boundary motion estimator 118 of FIG. In this region boundary motion estimation, as shown in FIG. 14, the search unit of motion vector search is performed using the boundary portions Γ ⁽¹⁾ , Γ ⁽²⁾ , Γ ⁽³⁾ , Let Γ ⁽⁴⁾ be the search unit. Here, a case where region boundary motion estimation is performed using Γ ⁽¹⁾ as the prediction target 1403 will be described.

First, as shown in FIG. 15A, in the encoded picture 1401, the boundary condition of the block boundary portion of Γ ⁽¹⁾ that is the prediction target 1403 is obtained. As the boundary condition of the block boundary part, it is desirable to use the inclination of the image signal at the boundary part as the boundary condition. Here, as an example, as shown in FIG. 15 (b), the virtual pixel at the boundary is represented by a cross, and from the pixel adjacent to the virtual pixel of the cross, the cross mark is shown as shown in FIG. 15 (c). DY and dV for specifying the inclination of the virtual pixel are calculated.

  In this way, the gradient information of the boundary portion, which is the boundary condition of the block boundary portion that is the prediction target 1403 in the encoded picture 1401, is obtained. In this embodiment, assuming that the obtained gradient information of the boundary portion is ideal gradient information, the boundary portion closest to the ideal gradient information is identified from the reference picture search range.

That is, in FIG. 14, first, a boundary portion of a block existing in the same spatial position of the reference picture 1402 as the boundary portion Γ ⁽¹⁾ of the block of the prediction target 1403 in the encoded picture 1401 is set as the reference position 1404. The inclination information 1404 is calculated and compared with the ideal inclination information. Here, the comparison method may use SAD used in normal motion estimation.

  After that, while moving the boundary portion in the reference picture 1402 to be compared with an accuracy of 1/2 pixel or 1/4 pixel, the inclination information of the boundary portion is calculated each time it is moved. The comparison with the inclination information is repeated, and the most suitable boundary portion 1406 within the search range 1405 of the reference picture 1402 is specified. Then, region boundary motion estimation is performed by obtaining a difference in spatial position between the boundary portion of the prediction target 1403 of the encoded picture 1401 and the matched boundary portion 1406 in the reference picture 1402 as boundary portion motion vector information 1407.

Then, as shown in FIG. 16, in the reference picture 1402 that matches the boundary portions Γ ⁽¹⁾ , Γ ⁽²⁾ , Γ ⁽³⁾ , Γ ⁽⁴⁾ of the block to be predicted in the encoded picture 1401 The inclination of the image signal, which is the boundary condition of the boundary portions 1601, 1602, 1603, and 1604, is considered as the inclination information of the image signal in the boundary portions 1601, 1602, 1603, and 1604 of the predicted picture 1605. Based on the specified inclination information of the image signal, a Poisson equation is applied to generate a prediction signal in the block, and the prediction signal is stored in the prediction block 1606. By performing such processing for each block in the prediction picture 1605 and generating a prediction signal, the region boundary motion compensation is completed.

  By performing region boundary motion estimation and region boundary motion compensation as described above, a predicted picture that is smoother at the block boundary than the predicted picture generated by motion estimation and motion compensation by block matching, which is a conventional method, is generated. Therefore, it becomes possible to generate a differential picture with good quality, which leads to improvement in coding efficiency.

  Next, the intra-regional differential motion estimation processing by the intra-regional differential motion estimator 121 provided in the moving image coding apparatus according to the present embodiment and the intra-regional differential motion compensation processing provided in the intra-regional differential motion compensator 122 will be described with reference to FIG. Will be described in detail with reference to FIG.

  FIG. 17 is a conceptual diagram showing how to obtain the reference difference information used in the intra-regional differential motion estimation process provided in the intra-regional differential motion estimator 121. In FIG. 17, in the intra-regional difference motion estimation of the present invention, the first predicted picture 1605 generated by the region boundary motion compensator 119 and the encoded picture 1401 to be encoded are acquired. Thereafter, using the acquired first predicted picture 1605 and encoded picture 1401, reference difference information 1702 of the block at the current processing target position is obtained.

Here, it is assumed that the prediction block 1606 in FIG. 17 is a block corresponding to the spatial position to be currently processed described with reference to FIG. In addition, the reference difference information 1702 in FIG. 17 is also a block corresponding to the spatial position currently being processed. This reference difference information 1702 includes signal information in a block to be predicted having boundary portions Γ ⁽¹⁾ , Γ ⁽²⁾ , Γ ⁽³⁾ , and Γ ⁽⁴⁾ in the encoded picture 1401 and the current processing target. The signal information in the prediction block 1606 corresponding to the spatial position to be identified is specified, and the difference 1701 between the two is generated.

  The reference difference information 1702 obtained in this way is used for motion vector search for specifying the difference component motion vector information 1802 in the reference picture 1302, as shown in FIG. The motion vector search for specifying the difference component motion vector information 1802 is a process for specifying a block 1801 having difference information most suitable for the obtained reference difference information 1702.

  Here, in FIG. 18, consider a process of determining whether or not the block 1803 that is the current search target within the search range 1305 of the reference picture 1302 is compatible with the reference difference information 1702. First, the difference between the signal component in the block 1803 that is the current search target and the prediction signal component of the corresponding region in the predicted picture is performed to generate search target difference information. Here, as shown in FIG. 18, the corresponding region in the prediction picture may be a prediction block 1606 that is a spatial position corresponding to the reference difference information 1702.

  Further, as shown in FIG. 19, the corresponding area in the predicted picture 1605 is the area 1901 in the predicted picture 1605 that is the spatial position corresponding to the current search target block 1803 in the search range 1305 of the reference picture 1302. And the difference from the predicted signal component in the specified region 1901 may be performed. Further, a configuration in which intra-regional differential motion estimation mode information is generated and supplied to the intra-regional differential motion compensator 122 and the entropy encoder 113 so that the methods of FIGS. 18 and 19 can be switched and used. Even better. The intra-region differential motion estimation mode information is easily obtained by flag information such as “0” for the region selection method from the predicted picture as shown in FIG. 18 and “1” for the region selection method from the prediction picture as shown in FIG. It is desirable to be able to specify the intra-regional differential motion estimation mode.

  In this way, according to the intra-region differential motion estimation mode, the corresponding region 1901 is identified from within the predicted picture 1605, and the search target difference information generated by performing the difference from the current search target region, The reference difference information 1702 is compared based on a predetermined determination method. Here, the predetermined determination method may use SAD as performed in normal motion estimation.

  By performing comparison based on this predetermined determination method and repeating the same processing while shifting the current search target position based on the predetermined motion accuracy until the most suitable region can be identified to the reference difference information 1702, As shown in FIGS. 18 and 19, intra-regional difference motion estimation is performed by identifying a finally adapted block 1801 and generating difference component motion vector information 1802 indicating the block 1801.

  Next, intra-regional differential motion compensation as shown in FIG. 20 or FIG. 21 is performed using the differential component motion vector information 1802 generated by the intra-regional differential motion estimation. FIG. 20 is a conceptual diagram showing the state of intra-region differential motion compensation when differential component motion vector information 1802 generated by intra-regional differential motion estimation as shown in FIG. 18 is obtained.

  In FIG. 20, a suitable block 1801 is identified from the reference picture 1302 based on the difference component motion vector information 1802. Thereafter, a prediction block 1606 at a spatial position corresponding to the prediction residual block 2001 in the prediction residual picture 2002 is identified from the prediction picture 1605, and the signal component of the identified prediction block 1606 and the matched block 1801 are A prediction residual block 2001 is generated by performing a difference calculation. By repeating similar processing, a prediction residual picture 2002 is generated and supplied to the second synthesizer 123 of FIG. The second synthesizer 123 synthesizes the prediction picture 1605 and the prediction residual picture 2002 as shown in FIG. 20 to generate a second prediction picture 2003.

  As an operation different from that in FIG. 20, in FIG. 21, a suitable block 1801 is specified from the reference picture 1302 based on the difference component motion vector information 1802. After that, unlike the case of FIG. 20, a region 2101 corresponding to the same spatial position as the adapted block 1801 is identified from the predicted picture 1605, and the difference between the signal component in this region 2101 and the adapted block 1801 is performed. Thus, a prediction residual block 2102 is generated. By repeating similar processing, a prediction residual picture 2103 is generated and supplied to the second synthesizer 123 of FIG. The second synthesizer 123 synthesizes the prediction picture 1605 and the prediction residual picture 2103 as shown in FIG. 21 to generate a second prediction picture 2104.

  Further, when the intra-regional difference motion estimation mode information is obtained, the processing of FIG. 20 is performed when the content of the flag information is “0”, and when the content of the flag information is “1”. More preferably, the process of FIG. 21 is performed.

  In the first embodiment of the moving picture coding apparatus according to the present invention described above, the discontinuous state between blocks is eliminated, and the second predicted picture closer to the coded picture including the continuous picture signal is obtained. As a result, the influence of motion estimation and motion compensation on the quality of the first difference picture generated by the difference between the subsequent second predicted picture and the encoded picture is eliminated. The effect of improving the efficiency of the orthogonal transform and quantization supplied with the difference picture is obtained.

  Further, according to the present embodiment, when performing region boundary motion estimation, the gradient of the image signal that is the boundary condition corresponding to each side of the rectangular region is separately searched for a motion vector for each side of the rectangular region, Even if the image signal included in the picture is deformed between the coded picture and the reference picture by generating each boundary motion vector information, separate boundary motion corresponding to each side is generated. By generating a first predictive picture with higher quality than the conventional motion compensation from the inclination of the image signal specified from the vector information, and further synthesizing a predictive residual picture to be closer to the encoded picture, Encoding efficiency can be improved.

  Next, an encoded bit stream generated by a computer that operates according to the first embodiment of the moving image encoding apparatus of the present invention or the first embodiment of the moving image encoding program of the present invention described later is used. A first embodiment of a moving picture decoding apparatus according to the present invention that acquires and decodes will be described below with reference to FIGS.

  FIG. 3 shows a block diagram of a first embodiment of a moving picture decoding apparatus according to the present invention. In the figure, the same components as those in FIG. 24 are denoted by the same reference numerals. As shown in FIG. 3, the moving picture decoding apparatus according to this embodiment is the first embodiment of the moving picture encoding apparatus shown in FIG. 1, or the first moving picture encoding program of the present invention described later. A computer-generated coded bitstream operating according to one embodiment receives an input code 301 as an input code 301 and decodes it by means of a switch 302, an input receiver 303, a demultiplexer 304, an entropy decoder. 305, inverse quantizer 306, inverse orthogonal transformer 307, first synthesizer 309, frame memory 311, decoding controller 313, regional boundary motion compensator 314, intra-regional differential motion compensator 315, and second synthesizer 316 At least. Further, as shown in FIG. 3, it is desirable to have a configuration including an intra predictor 308 and a deblock filter 310.

  The switch 302 performs decoding control on one of the third predicted picture from the intra predictor 308 supplied to the terminal a and the second predicted picture from the second synthesizer 316 supplied to the terminal b. And a function of supplying the selected predicted picture to the first synthesizer 309.

  The input receiver 303 has a function of acquiring packet information obtained by packetizing an encoded bit stream (input code 301) supplied from a predetermined transmission path or a predetermined recording medium by predetermined packet processing, and the acquired packet By performing packet combining processing from information, it has a function of generating an encoded bit stream and a function of supplying the generated encoded bit stream to the demultiplexer 304.

  The demultiplexer 304 obtains an encoded bitstream from the input receiver 303, performs a demultiplexing of the encoded bitstream based on a predetermined syntax structure, and generates a coded bit string A function of supplying the encoded bit string to the entropy decoder 305.

  The entropy decoder 305 acquires at least encoded information, boundary motion vector information, and difference component motion vector information by acquiring a coded bit string from the demultiplexer 304 and performing predetermined entropy decoding. It has a function. Here, it is desirable that the entropy decoder 305 has a function of acquiring parameter information and the like necessary for configuring another predetermined syntax structure by performing entropy decoding on the encoded bit string.

  Parameter information necessary to configure a predetermined syntax structure includes macroblock information for specifying various states of the macroblock, quantization parameter information used when performing quantization and inverse quantization, and an intra prediction method It is further preferable that intra prediction mode information for specifying the frame order information, frame order information for specifying the reference order of the reference image frames, and the like. In addition, the entropy decoder 305 supplies the obtained post-quantization information to the inverse quantizer 306, and the boundary motion vector information is supplied to the region boundary motion compensator 314, and the difference component motion vector information is supplied to the intra-region differential motion compensation At least a function of supplying the unit 315.

  The inverse quantizer 306 obtains post-quantization information from the entropy decoder 305, performs predetermined inverse quantization based on a predetermined quantization parameter, generates post-quantization information, and performs inverse orthogonal It has a function of supplying to the converter 307. The inverse orthogonal transformer 307 acquires the post-inverse quantization information from the inverse quantizer 306 and performs a predetermined inverse orthogonal transform to generate a decoded differential picture, and the generated decoded differential picture is converted into a first And a function to supply to the combiner 309.

  The first synthesizer 309 acquires a decoded differential picture from the inverse orthogonal transformer 307, generates a decoded picture by synthesizing the acquired decoded differential picture and the predicted picture, and a function of acquiring the predicted picture from the switch 302. And a function of supplying the generated decoded picture to the intra predictor 308 and the deblocking filter 310.

  The frame memory 311 has a function of acquiring and storing a decoded picture from each unit in order to hold the decoded picture as a reference picture. In FIG. 3, it is desirable that the decoded picture after the deblocking filter process is acquired from the deblocking filter 310. Further, the frame memory 311 has a function of supplying the requested reference picture to each unit of the decoding device as necessary. In FIG. 3, the reference picture is preferably configured to be supplied to the intra predictor 308, the regional boundary motion compensator 314, and the intra-regional differential motion compensator 315. The frame memory 311 has a function of supplying a decoded picture as an output code 312 in response to a request from the output display device. In order to control the operation of the decoding apparatus of the present invention, the decoding controller 313 has a function of supplying various parameter information necessary for decoding to each unit constituting the decoding apparatus, and input / output of each unit. And at least a function of performing switching control of the switch 302.

  The region boundary motion compensator 314 has a function of acquiring boundary portion motion vector information from the entropy decoder 305 and acquiring a reference picture from the frame memory 311. In addition, the region boundary motion compensator 314 specifies, from the reference picture, the inclination of the image signal of each side necessary for generating the estimation signal in the block based on the acquired boundary portion motion vector information. The prediction signal in the block is generated by applying the slope of the image signal of each specified side as a boundary condition to the Poisson equation, and the prediction signal of each block in the picture is generated to generate the first prediction picture. Thus, it has a function of performing region boundary motion compensation and a function of supplying the generated first predicted picture to the second synthesizer 316 and the intra-regional difference motion compensator 315.

  The intra-regional difference motion compensator 315 acquires a corresponding reference picture from the frame memory 311 and a first predicted picture from the region boundary motion compensator 314, and acquires difference component motion vector information from the entropy decoder 305. Have Further, the intra-region difference motion compensator 315 identifies a corresponding block region from the reference picture based on the acquired difference component motion vector information, and the signal component in the block region in the identified reference picture, A function of generating a prediction residual picture that is a differential prediction picture from a difference from a signal component in a corresponding rectangular area in the prediction picture, and a function of supplying the generated prediction residual picture to the second synthesizer 316 Have.

  The second synthesizer 316 acquires the first predicted picture from the regional boundary motion compensator 314, acquires the prediction residual picture from the intra-regional differential motion compensator 315, and the acquired first predicted picture and prediction It has a function of combining the residual picture and generating a second predicted picture, and a function of supplying the generated second predicted picture to the terminal b side of the switch 302. The intra predictor 308 obtains a decoded picture from the first synthesizer 309 and performs a predetermined intra prediction to generate a third predicted picture, and the generated third predicted picture is connected to the terminal of the switch 302. a. Note that the configuration may be such that a picture used for intra prediction is acquired from the frame memory 311.

  The deblock filter 310 acquires the decoded picture from the first combiner 309, performs a predetermined deblock filter process, and then supplies the decoded picture to the frame memory 311 to store the decoded picture as a reference memory. Has the function of

  Next, the operation of the moving picture decoding apparatus shown in FIG. 3 will be described according to the flowchart of FIG. First, the input receiver 303 is generated by a computer that operates according to the first embodiment of the moving picture coding apparatus shown in FIG. 1 or the first embodiment of the moving picture coding program of the present invention described later. The encoded bit stream is received from a predetermined transmission path or a predetermined recording medium as an input code 301, and a predetermined packet combining process is performed from the received input code 301 to reconstruct the encoded bit stream. (Step S201). Thereafter, the input receiver 303 supplies the reconstructed encoded bitstream to the demultiplexer 304.

  The demultiplexer 304 acquires the encoded bitstream reconstructed by the input receiver 303, and performs demultiplexing of encoded information on the acquired encoded bitstream based on a predetermined syntax structure ( In step S202), the entropy decoder 305 is notified of the encoded bit string that is the information after demultiplexing.

  The entropy decoder 305 acquires the encoded bit string from the demultiplexer 304, performs predetermined entropy decoding (step S203), post-quantization information, boundary motion vector information, difference component motion vector information, and Parameter information and the like necessary for constructing other predetermined syntax structures are generated. Thereafter, the entropy decoder 305 supplies at least post-quantization information to the inverse quantizer 306, boundary region motion vector information to the region boundary motion compensator 314, and difference component motion vector information to intra region differential motion compensation. To the container 315.

  Thereafter, the decoding controller 313 determines whether the current decoding mode of the decoding apparatus is intra decoding or not intra decoding (step S204). In the case of intra decoding, the decoding controller 313 supplies parameter information for performing intra decoding to each unit of the decoding apparatus, and connects the switch 302 to the terminal a, so that the intra unit is connected. Prepare for decryption. Then, predetermined intra prediction is performed by the intra predictor 308 (step S205). Thereafter, the process proceeds to step S208.

  On the other hand, when it is not intra decoding, the decoding controller 313 supplies parameter information for performing inter decoding to each unit, and prepares for performing inter decoding by connecting the switch 302 to the terminal b. do. Thereafter, the region boundary motion compensator 314 obtains the boundary motion vector information from the entropy decoder 305 and the corresponding reference picture from the frame memory 311 and performs predetermined region boundary motion compensation (step S206). The predicted picture is generated. The region boundary motion compensator 314 supplies the generated first predicted picture to the intra-region difference motion compensator 315 and the second synthesizer 316, respectively.

  Further, the intra-region differential motion compensator 315 acquires the difference component motion vector information from the entropy decoder 305, acquires the corresponding reference picture from the frame memory 311, and further receives the first prediction from the region boundary motion compensator 314. Based on these, a predetermined intra-regional differential motion compensation is performed (step S207), and a prediction residual picture is generated and supplied to the second synthesizer 316. The second combiner 316 combines the first predicted picture from the region boundary motion compensator 314 and the prediction residual picture from the intra-region difference motion compensator 315 to generate a second predicted picture. . This second predicted picture is a synthesized predicted picture that is synthesized with the decoded differential picture by a first synthesizer 309 described later. The second synthesizer 316 supplies the generated second predicted picture to the terminal b side of the switch 302. Thereafter, the process proceeds to step S208.

  When the prediction picture preparation is completed, the inverse quantizer 306 obtains post-quantization information from the entropy decoder 305, performs predetermined inverse quantization based on a predetermined quantization parameter (step S208), and performs inverse processing. Generate post-quantization information. The generated dequantized information is supplied to the inverse orthogonal transformer 307. The inverse orthogonal transformer 307 acquires information after inverse quantization from the inverse quantizer 306, performs predetermined inverse orthogonal transform (step S209), and generates a decoded difference picture. The generated decoded differential picture is supplied to the first combiner 309.

  The first synthesizer 309 synthesizes the decoded differential picture from the inverse orthogonal transformer 307 and the intra prediction picture or the second prediction picture (composition prediction picture) from the switch 302 (step S210). Generate. The generated decoded picture is supplied to the deblocking filter 310. Also, when the decoding mode is intra decoding, the decoded picture is further supplied to the intra predictor 308.

  The deblocking filter 310 acquires a decoded picture from the first combiner 309 and performs a predetermined deblocking filter process for removing block distortion (step S211). The decoded picture generated after the deblocking filter processing is stored in the frame memory 311 as a reference picture, thereby preparing a reference picture for the next encoding (step S212).

  In the frame memory 311, the decoded picture after the deblocking filter processing is acquired from the deblocking filter 310 and stored as a reference picture. Also, reference pictures are supplied to the regional boundary motion compensator 314, the intra-regional differential motion compensator 315, and the intra predictor 308 as necessary. Here, it is desirable that the frame memory 311 accumulates at least one reference picture and can supply necessary reference pictures according to the request of each component. The frame memory 311 outputs a decoded image frame in accordance with the display timing of the decoded picture (step S213), and supplies the decoded picture necessary for display as an output code 312 to an external output display.

  After that, if there is a further encoded bit stream to be decoded and decoding is necessary, the decoding process is continued by proceeding to step S201. If decoding is not necessary, a series of decoding is performed. The process is terminated.

(Second Embodiment)
Next, a second embodiment of the moving picture coding apparatus and moving picture decoding apparatus according to the present invention will be described with reference to FIGS. FIG. 5 shows a block diagram of a second embodiment of the moving picture coding apparatus according to the present invention. In the figure, the same components as those in FIG. As shown in FIG. 5, the moving picture coding apparatus according to the present embodiment has a switch 501, a motion estimator 502, a motion compensator 503, and a difference determiner 504 in addition to the configuration of the moving picture coding apparatus in FIG. Is further provided.

  The switch 501 is supplied with the third predicted picture generated by the intra predictor 105 as an input to the terminal a, and is supplied with the fourth predicted picture generated by the motion compensator 503 as an input to the terminal b. A function of selecting one of the fourth predicted pictures according to the control of the encoding controller 117 and supplying the selected third or fourth predicted picture to the terminal b of the switch 103 and the second subtractor 124. Have.

  The motion estimator 502 obtains a coded picture that is a coding target of a moving image that is the input code 101 and a reference picture that is a target from the frame memory 104, and performs a predetermined motion estimation to obtain a motion vector. It has a function to generate information. In the above-described predetermined motion estimation, similarly to motion estimation such as general AVC, a motion vector search is performed by block matching in a reference picture in each block area in an encoded picture, and a motion vector up to a matching block area is obtained. Generate information. The motion estimator 502 has a function of supplying the generated motion vector information to the motion compensator 503 and the entropy encoder 113.

  The motion compensator 503 specifies a corresponding block region from the corresponding reference picture acquired from the frame memory 104 based on the motion vector information acquired from the motion estimator 502, and generates a fourth predicted picture; A function of supplying the generated fourth predicted picture to the terminal b of the switch 501;

  The difference determiner 504 includes a second difference picture generated by performing a difference operation between the third or fourth predicted picture and the encoded picture that is the input code 101 in the second differencer 124, and the first difference The unit 120 has a function of acquiring at least a first difference picture generated by performing a difference operation between the second predicted picture and the encoded picture that is the input code 101. Further, the difference determiner 504 further acquires boundary portion motion vector information from the region boundary motion compensator 118, difference component motion vector information from the intra-region difference motion estimator 121, and motion vector information from the motion estimator 502. If it has a function, it will become a better structure.

  Further, the difference determiner 504 has a function of determining which of the acquired first difference picture or second difference picture is to be used based on a predetermined determination criterion. Here, the predetermined criterion is a criterion for comparing the acquired information amount of the first difference picture and the information amount of the second difference picture and selecting the difference picture having the smaller information amount. It is good to be. In addition, the predetermined criterion is that the information amount of the boundary motion vector information, the information amount of the difference component motion vector information, and the information amount of the motion vector information are further used to obtain the first difference picture of the acquired first difference picture. The information amount, the information amount of the boundary motion vector information, and the difference component motion vector information are compared with the information amount of the second difference picture and the information amount of the motion vector information, and the difference picture with the smaller information amount is selected. If it is such a determination standard, it will become a still more favorable structure. Further, the difference determiner 504 has a function of supplying difference determination information indicating which of the first difference picture and the second difference picture is selected to the encoding controller 117 and the entropy encoder 113.

  The encoding controller 117 acquires the difference determination information from the difference determiner 504, and controls the switching of the switch 102 based on the determination result indicated by the acquired difference determination information, so that the difference picture supplied to the orthogonal transformer 107 And the switching control of the switch 103 based on the determination result indicated by the acquired difference determination information, the predicted picture used when generating the selected difference picture is converted into the first synthesizer 111. Has a function of performing prediction picture switching control so as to be supplied to. The encoding controller 117 has a function of performing switching control of the switch 501 in accordance with the encoding mode. The selection result of the switch 501 is configured to be supplied to the terminal b of the switch 103.

  The intra predictor 105 has a function of supplying the generated third predicted picture to the terminal a of the switch 501. In addition to the function of the entropy encoder 113 of FIG. 1, the entropy encoder 113 further acquires motion vector information from the motion estimator 502 and difference determination information from the difference determiner 504, and performs predetermined entropy encoding. Has the function to perform.

  The second subtractor 124 has a function of acquiring an encoded picture that is an encoding target of a moving image that is the input code 101 and acquiring a third predicted picture or a fourth predicted picture from the switch 501. Here, the third predicted picture is generated by the intra predictor 105, and the fourth predicted picture is generated by the motion compensator 503. Further, the second subtractor 124 performs a difference operation between the acquired encoded picture and the predicted picture, thereby generating a second difference picture, and the generated second difference picture is connected to the terminal b of the switch 102. And the difference determination unit 504.

  Other functions included in the configuration of the moving picture encoding apparatus of the present embodiment are the same as the functions of the respective sections included in the first embodiment of the moving picture encoding apparatus shown in FIG. For this reason, the same reference numerals are given here, and the description is omitted to avoid duplication.

  Next, the operation of the second embodiment of the video encoding apparatus shown in FIG. 5 will be described with reference to the flowchart of FIG. First, as an input code 101, a frame or a field that is an encoding target of a moving image is prepared as an encoded picture (step S301). The encoding controller 117 determines whether the current encoding mode of the encoding apparatus is intra encoding or not intra encoding (step S302).

  In the case of intra coding, the coding controller 117 supplies parameter information for performing intra coding to each unit of the coding apparatus, connects the switch 102 to the terminal b, and switches the switch 103 to By connecting to the terminal b and connecting the switch 501 to the terminal a, preparation for intra coding is performed. Thereafter, predetermined intra prediction is performed by the intra predictor 105 (step S303). Thereafter, the process proceeds to step S310.

  On the other hand, when the encoding is not intra encoding, the encoding controller 117 supplies parameter information for performing inter encoding to each unit, disconnects the switch 102, disconnects the switch 103, and connects the switch 501 to the terminal b. By doing so, it prepares for performing inter coding. Thereafter, the region boundary motion estimator 118 obtains the encoded picture prepared as the input code 101 and the reference picture stored in the frame memory 104, and performs predetermined region boundary motion estimation based on these (step In step S <b> 304, boundary motion vector information is generated, and the generated boundary motion vector information is supplied to the region boundary motion compensator 119 and also supplied to the entropy encoder 113.

  The region boundary motion compensator 119 performs predetermined region boundary motion compensation based on the boundary motion vector information acquired from the region boundary motion estimator 118 and the corresponding reference picture acquired from the frame memory 104 (step S305). ), A first prediction picture is generated, and the generated first prediction picture is supplied to the second synthesizer 123, the intra-regional difference motion estimator 121, and the intra-regional difference motion compensator 122, respectively.

  The intra-region difference motion estimator 121 acquires the encoded picture to be encoded as the input code 101, and also obtains the corresponding reference picture from the frame memory 104 and the first predicted picture from the region boundary motion compensator 119, respectively. Obtaining and performing intra-regional differential motion estimation based on each acquired picture (step S306), generating differential component motion vector information, and using the generated differential component motion vector information as intra-regional differential motion compensator 122 and entropy code To the generator 113.

  The intra-region difference motion compensator 122 is a corresponding reference picture acquired from the frame memory 104, a first predicted picture acquired from the region boundary motion compensator 119, and a difference component motion acquired from the intra-region difference motion estimator 121. Based on the vector information, intra-region difference motion compensation is performed (step S307), a prediction residual picture is generated, and the generated prediction residual picture is supplied to the second synthesizer 123. After that, the second synthesizer 123 generates a second predicted picture based on the first predicted picture acquired from the regional boundary motion compensator 119 and the predicted residual picture acquired from the intra-regional differential motion compensator 122. The generated second predicted picture is supplied to the terminal a of the switch 103 and the first differentiator 120, respectively.

  Similarly, the motion estimator 502 acquires a coded picture prepared as the input code 101 and a reference picture stored in the frame memory 104, performs predetermined motion estimation (step S308), and obtains motion vector information. The generated motion vector information is supplied to the motion compensator 503 and also supplied to the entropy encoder 113.

  The motion compensator 503 performs predetermined motion compensation based on the motion vector information acquired from the motion estimator 502 and the corresponding reference picture acquired from the frame memory 104 (step S309), and obtains the fourth predicted picture. The generated fourth prediction picture is supplied to the terminal b of the switch 501. Thereafter, the process proceeds to step S310.

  Subsequently, in step S310, a difference picture is generated by performing a difference operation between the prediction picture generated by the intra predictor 105 or the second synthesizer 123 and the motion compensator 503 and the encoded picture that is the input code 101. To do. Here, when the coding mode is intra coding, since the switch 501 is connected to the terminal a, the second subtractor 124 includes the coded picture that is the input code 101 and the intra predictor 105. Is obtained, and a difference calculation is performed between the obtained encoded picture and the third predicted picture to generate a second difference picture, and a difference determination is made between the terminal b of the switch 102 and the difference determination. To the vessel 504.

  On the other hand, when the coding mode is inter coding, since the switch 501 is connected to the terminal b, the second subtractor 124 includes the coded picture as the input code and the motion compensator 503. Is obtained, and a difference calculation between the obtained encoded picture and the fourth predicted picture is performed to generate a second difference picture, and a difference determination is made between the terminal b of the switch 102 and the difference determination. To the vessel 504.

  In the first subtractor 120, the encoded picture that is the input code 101 and the second predicted picture generated by the second synthesizer 123 are acquired, and the acquired encoded picture and the second predicted picture are obtained. The first difference picture is generated by performing the difference calculation, and is supplied to the terminal a of the switch 102 and the difference determination unit 504.

  The difference determiner 504 compares the acquired information amount of the first difference picture with the information amount of the second difference picture, and selects a difference picture with a smaller information amount based on a predetermined determination criterion. Then, it is determined which one of the acquired first difference picture and second acquired difference picture is used (step S311). Further, the difference determiner 504 supplies difference determination information indicating which one of the first difference picture and the second difference picture is adopted to the encoding controller 117 and the entropy encoder 113, respectively.

  When difference determination information is supplied from the difference determiner 504 to the encoding controller 117, the encoding controller 117 switches the switch 102 and the switch 103 to control encoding. Here, when the encoding controller 117 specifies from the difference determination information that the first difference picture has been adopted, the encoding controller 117 connects the switch 102 to the terminal a and the switch 103 to the terminal a. When it is determined from the difference determination information that the second difference picture has been adopted, the switch 102 is connected to the terminal b and the switch 103 is connected to the terminal b. In this way, the encoding controller 117 controls the switch 102 so that the difference picture adopted by the difference judgment unit 504 is supplied to the orthogonal transformer 107, and the difference picture adopted by the difference judgment unit 504. The switch 103 is controlled so as to supply the corresponding predictive picture used when generating the first synthesizer 111 to the first synthesizer 111.

  The orthogonal transformer 107 performs predetermined orthogonal transformation on the difference picture selected by the switch 102 and adopted by the difference determiner 504 (step S312), and generates orthogonal transformation coefficient information (here, DCT coefficient information). Then, the generated DCT coefficient information is supplied to the quantizer 108. The quantizer 108 acquires DCT coefficient information from the orthogonal transformer 107, performs predetermined quantization based on a predetermined quantization parameter (step S313), and generates post-quantization information. The generated post-quantization information is supplied to the inverse quantizer 109 and the entropy encoder 113.

  The inverse quantizer 109 acquires post-quantization information from the quantizer 108, performs predetermined inverse quantization based on a predetermined quantization parameter (step S314), generates information after inverse quantization, and performs inverse processing. This is supplied to the orthogonal transformer 110. The inverse orthogonal transformer 110 acquires information after inverse quantization from the inverse quantizer 109, performs predetermined inverse orthogonal transform (step S315), generates a decoded differential picture, and supplies the decoded differential picture to the first combiner 111.

  The first combiner 111 combines the decoded differential picture acquired from the inverse orthogonal transformer 110 and the predicted picture acquired from the switch 103 (step S316), and generates a decoded picture that is a locally decoded image signal. The decoded picture is supplied to the deblock filter 112. When the encoding mode is intra encoding, the decoded picture is further supplied to the intra predictor 105.

  The deblocking filter 112 performs a predetermined deblocking filter process on the decoded picture acquired from the first combiner 111 (step S317), generates a decoded picture after the deblocking filter process, and uses the frame memory as a reference picture. By storing in 104, a reference picture for the next encoding is prepared (step S318).

  Thereafter, in order to output a series of encoding results, the entropy encoder 113 includes at least quantized information from the quantizer 108, boundary motion vector information from the region boundary motion estimator 118, and intra-regional differential motion estimator. The difference component motion vector information is acquired from 121, the motion vector information is acquired from the motion estimator 502, and the difference determination information is acquired from the difference determiner 504, and predetermined entropy encoding is performed (step S319) to generate an encoded bit string. Here, the entropy encoder 113 may further acquire various parameter information used at the time of encoding from each unit of the encoding device and perform predetermined entropy encoding. The generated encoded bit string is supplied to the multiplexer 114.

  The multiplexer 114 multiplexes the encoded bit string acquired from the entropy encoder 113 based on a predetermined syntax structure (step S320), generates an encoded bit stream, and converts the encoded bit stream to This is supplied to the output transmitter 115. The output transmitter 115 performs predetermined packetization processing or the like on the multiplexed bit stream acquired from the multiplexing unit 114, and then outputs and transmits the encoding result to a predetermined transmission path or a predetermined recording medium (step). S321). By performing the processes of steps S301 to S321 described above, the encoding process of the encoded picture that is the encoding target at a certain point in time is completed in the moving image encoding apparatus of the present embodiment.

  In the present embodiment, the difference determiner 504 switches the difference pictures generated so as to reduce the amount of information to be generated, so that the coding efficiency of the entire coded bitstream can be improved.

  Next, an encoded bit stream generated by a computer that operates according to the second embodiment of the moving image encoding apparatus of the present invention or the second embodiment of the moving image encoding program of the present invention described later is obtained. A second embodiment of the moving picture decoding apparatus according to the present invention that performs acquisition and decoding will be described below with reference to FIGS.

  FIG. 7 shows a block diagram of a second embodiment of the moving picture decoding apparatus according to the present invention. In the figure, the same components as in FIG. As shown in FIG. 7, the moving picture decoding apparatus of this embodiment is a second embodiment of the moving picture encoding apparatus shown in FIG. 5, or a moving picture encoding program of the present invention to be described later. In order to receive an encoded code stream generated by a computer operating according to the second embodiment as an input code 301 and decode it, the first embodiment of the video decoding apparatus shown in FIG. In addition to the configuration, a motion compensator 317 and a switch 318 are further provided.

  The switch 318 is supplied with the second predicted picture from the second synthesizer 316 to the terminal b and supplied with the fourth predicted picture from the motion compensator 317 to the terminal a. The switch 318 has a function of supplying the supplied second predicted picture and fourth predicted picture to the terminal b of the switch 302 by switching according to the request of the decoding controller 313.

  The motion compensator 317 obtains motion vector information from the entropy decoder 305 and also obtains a corresponding reference picture from the frame memory 311 and a corresponding block area from the reference picture based on the obtained motion vector information. A function for generating a fourth predicted picture, and a function for supplying the generated fourth predicted picture to the terminal a of the switch 318.

  In the switch 302, the third predicted picture generated by the intra predictor 308 is supplied to the terminal a, and the second predicted picture or the fourth predicted picture is supplied to the terminal b by switching control of the switch 318. . In addition, the switch 302 has a function of performing switching control according to a request of the decoding controller 313 for each prediction picture supplied to the terminal a and the terminal b, and a prediction picture selected by performing the switching control. 1 synthesizer 309.

  The entropy decoder 305 has a function of acquiring motion vector information and difference determination information by performing entropy decoding on the encoded bit sequence acquired from the demultiplexer 304, and motion compensation of the acquired motion vector information. And a function of supplying the obtained difference determination information to the decoding controller 313.

  The decoding controller 313 acquires the difference determination information from the entropy decoder 305, and the predicted pictures used when generating the difference picture from the acquired difference determination information are the second predicted picture and the fourth predicted picture. And a function of performing switching control of the switch 318 based on the result specified from the difference determination information. Here, when the decoding controller 313 identifies that the prediction picture used when generating the difference picture from the difference determination information is the second prediction picture, the decoding controller 313 connects the switch 318 to the terminal b. To do. In addition, when it is determined that the predicted picture used when generating the difference picture from the difference determination information is the fourth predicted picture, the switch 318 is connected to the terminal a. The predicted picture selected by the switch 318 is supplied to the terminal b of the switch 302.

  In addition, the decoding controller 313 has a function of performing switching control of the switch 302 according to the decoding mode. Here, when the decoding mode is intra decoding, the switch 302 is connected to the terminal a, and when the decoding mode is not intra decoding, the switch 302 is connected to the terminal b, thereby switching. Take control. The second synthesizer 316 has a function of supplying the generated second predicted picture to the terminal b of the switch 318. Other functions included in the configuration of the moving picture decoding apparatus according to the present embodiment are the same as the functions of the respective sections included in the first embodiment of the moving picture decoding apparatus shown in FIG. Therefore, here, in order to avoid duplication, the same reference numerals are given and the description is omitted.

  Next, the operation of the second embodiment of the moving picture decoding apparatus of the present invention shown in FIG. 7 will be described with reference to the flowchart of FIG. First, the input receiver 303 is generated by a computer that operates according to the second embodiment of the moving picture coding apparatus shown in FIG. 5 or the second embodiment of the moving picture coding program of the present invention described later. The encoded bit stream is received from a predetermined transmission path or a predetermined recording medium as an input code 301, and a predetermined packet combining process is performed from the received input code 301 to reconstruct the encoded bit stream. (Step S401). Thereafter, the input receiver 303 supplies the reconstructed encoded bitstream to the demultiplexer 304.

  The demultiplexer 304 acquires the encoded bitstream reconstructed by the input receiver 303, and performs demultiplexing of encoded information on the acquired encoded bitstream based on a predetermined syntax structure ( In step S402), the entropy decoder 305 is notified of an encoded bit string that is information after demultiplexing.

  The entropy decoder 305 acquires the encoded bit string from the demultiplexer 304, performs predetermined entropy decoding (step S403), and performs post-quantization information, boundary motion vector information, difference component motion vector information, motion Generates vector information, difference determination information, and other parameter information necessary for constructing a predetermined syntax structure, and at least post-quantization information is input to the inverse quantizer 306 and boundary motion vector information is input to the region boundary motion. The compensator 314 is supplied with the difference component motion vector information to the intra-regional difference motion compensator 315, the motion vector information to the motion compensator 317, and the difference determination information to the decoding controller 313.

  After that, the decoding controller 313 determines whether the current decoding mode of the decoding device is intra decoding or not intra decoding (step S404). In the case of intra decoding, the decoding controller 313 supplies parameter information for performing intra decoding to each unit of the decoding apparatus, and connects the switch 302 to the terminal a, so that the intra unit is connected. Prepare for decryption. Thereafter, predetermined intra prediction is performed by the intra predictor 308 (step S405). Thereafter, the process proceeds to step S410. On the other hand, when it is not intra decoding, the decoding controller 313 supplies parameter information for performing inter decoding to each unit, and prepares for performing inter decoding by connecting the switch 302 to the terminal b. do.

  Next, the decoding controller 313 determines whether it is the region boundary mode or not the region boundary mode based on the difference determination information acquired from the entropy decoder 305 (step S406). If it is determined that the region boundary mode is selected, the decoding controller 313 connects the switch 318 to the terminal b so as to select the second predicted picture generated by the second synthesizer 316, and then selects the region boundary. The motion compensator 314 performs predetermined region boundary motion compensation based on the boundary portion motion vector information acquired from the entropy decoder 305 and the corresponding reference picture acquired from the frame memory 311 (step S407). 1 predicted picture is generated, and the first predicted picture is supplied to the second synthesizer 316 and the intra-region difference motion compensator 315, respectively.

  Subsequently, the intra-region difference motion compensator 315 receives the difference component motion vector information acquired from the entropy decoder 305, the corresponding reference picture acquired from the frame memory 311, and the first acquired from the region boundary motion compensator 314. Based on the predicted picture, predetermined intra-regional differential motion compensation is performed (step S408), a prediction residual picture is generated, and the prediction residual picture is supplied to the second synthesizer 316. The second synthesizer 316 generates a second predicted picture based on the first predicted picture acquired from the region boundary motion compensator 314 and the prediction residual picture acquired from the intra-region difference motion compensator 315, The second predicted picture is supplied to the terminal b of the switch 318. Thereafter, the process proceeds to step S410.

  On the other hand, if the decoding controller 313 determines in step S406 that it is not the region boundary mode, the decoding controller 313 connects the switch 318 to the terminal a so as to select the fourth predicted picture generated by the motion compensator 317, and The motion compensator 317 performs predetermined motion compensation based on the motion vector information acquired from the entropy decoder 305 and the corresponding reference picture acquired from the frame memory 311 (step S409), and the fourth A predicted picture is generated, and the fourth predicted picture is supplied to the terminal a of the switch 318. Thereafter, the process proceeds to step S410.

  When the prediction picture preparation is completed by the processing in step S405, S408, or S409, the inverse quantizer 306 acquires post-quantization information from the entropy decoder 305, and performs predetermined inverse quantum based on a predetermined quantization parameter. (Step S410). The information after inverse quantization generated in this way is supplied to the inverse orthogonal transformer 307. The inverse orthogonal transformer 307 acquires the post-inverse quantization information from the inverse quantizer 306, performs predetermined inverse orthogonal transform (step S411), generates a decoded difference picture, and generates the decoded difference picture as the first combiner. 309.

  The first synthesizer 309 synthesizes the decoded differential picture acquired from the inverse orthogonal transformer 307 and the predicted picture acquired from the switch 302 (step S412), generates a decoded picture, and outputs the decoded picture to the deblock filter 310. To supply. When the decoding mode is intra decoding, the decoded picture is further supplied to the intra predictor 308.

  The deblocking filter 310 performs a predetermined deblocking filter process on the decoded picture acquired from the first combiner 309 (step S413). The decoded picture after the deblocking filter process is stored in the frame memory 311 as a reference picture, so that it is prepared as a reference picture for the next encoding (step S414).

  In the frame memory 311, the decoded picture after the deblocking filter processing is acquired and accumulated as a reference picture from the deblocking filter 310, and the region boundary motion compensation 314, the intraregional difference motion compensator 315, and the motion compensator 317 are acquired as necessary. The reference picture is supplied to the intra predictor 308. Here, it is desirable that the frame memory 311 accumulates at least one reference picture and can supply necessary reference pictures according to the request of each component. The frame memory 311 outputs a decoded image frame in accordance with the display timing of the decoded picture (step S415), and supplies the decoded picture necessary for display as an output code 312 to an external output display.

  After that, if there is a further encoded bit stream to be decoded and decoding is necessary, the decoding process is continued by proceeding to step S401. If there is no need for decoding, a series of decoding is performed. The process is terminated.

(Program embodiment)
The present invention relates to the central processing control apparatus which is a computer shown in FIGS. 9 to 12 with the functions of the moving picture encoding apparatus shown in FIGS. 1 and 5 and the functions of the moving picture decoding apparatus shown in FIGS. A program for realizing the program 903 is included, and the embodiments thereof will be described below.

  FIG. 9 is a block diagram showing an example of an information processing apparatus that operates according to the first embodiment of the moving picture coding program of the present invention. In the figure, an information processing apparatus 900 is operated by an input device 901 for inputting various types of information, an output unit 902 for outputting various types of information, and the moving picture encoding program of the first embodiment. A central processing control device 903, an external storage device 904, a temporary storage device 905 used as a work area for arithmetic processing by the central processing control device 903, and a communication device 906 for communicating with the outside are bidirectional. The bus 907 is connected.

  The central processing control apparatus 903 distributes the moving picture coding program of the first embodiment from a recording medium or via a communication network and takes it in by the communication apparatus 906, and the intra processing predictor 105 shown in FIG. Corresponding intra prediction means 908, first difference means 909 corresponding to the first differencer 120, orthogonal transform means 910 corresponding to the orthogonal transformer 107, quantization means 911 corresponding to the quantizer 108, and inverse quantizer 109 Inverse quantization means 912 corresponding to, inverse orthogonal transform means 913 corresponding to the inverse orthogonal transformer 110, first synthesis means 914 corresponding to the first combiner 111, deblock filter means 915 corresponding to the deblock filter 112, Entropy encoding means 916 corresponding to the entropy encoder 113 and multiplexing means 91 corresponding to the multiplexer 114 Output transmission means 918 corresponding to the output transmitter 115, encoding control means 919 corresponding to the encoding controller 117, region boundary motion estimation means 920 corresponding to the region boundary motion estimator 118, and region boundary motion compensator 119 Corresponding to the corresponding regional boundary motion compensation unit 921, the second differential unit 922 corresponding to the second differential unit 124, the intra-regional differential motion estimation unit 923 corresponding to the intra-regional differential motion estimator 121, and the intra-regional differential motion compensator 122 1 has at least the functions of the intra-regional differential motion compensation unit 924 and the second synthesis unit 925 corresponding to the second synthesizer 123. For example, a temporary storage device 905 is used as a storage unit corresponding to the frame memory 104 shown in FIG. By using this, the same operation as that of the moving picture encoding apparatus shown in FIG. 1 is executed by software processing.

  FIG. 10 is a block diagram showing an example of an information processing apparatus that operates according to the first embodiment of the moving picture decoding program of the present invention. In the figure, the same components as those in FIG. 9 are denoted by the same reference numerals, and the description thereof is omitted. The basic configuration of the information processing apparatus 900 illustrated in FIG. 10 is the same as that illustrated in FIG.

  However, the central processing control apparatus 903 shown in FIG. 10 receives the moving picture decoding program according to the first embodiment of the present invention from a recording medium or via a communication network and is captured by the communication apparatus 906. The input receiving unit 1001 corresponding to the input receiver 303 shown in FIG. 3, the demultiplexing unit 1002 corresponding to the demultiplexing unit 304, the entropy decoding unit 1003 corresponding to the entropy decoding unit 305, and the inverse quantizer 306 Inverse quantization means 1004 corresponding, inverse orthogonal transform means 1005 equivalent to inverse orthogonal transformer 307, intra prediction means 1006 equivalent to intra predictor 308, first synthesis means 1007 equivalent to first combiner 309, deblocking Deblock filter means 1008 corresponding to the filter 310 and decoding corresponding to the decoding controller 313 The control unit 1009, the region boundary motion compensation unit 1010 corresponding to the region boundary motion compensator 314, the intra-regional difference motion compensation unit 1011 corresponding to the intra-regional difference motion compensator 315, and the second synthesis unit corresponding to the second synthesizer 316. By using, for example, a temporary storage device 905 as storage means corresponding to the frame memory 311 shown in FIG. 3 at least having the functions 1012, the same operation as the moving picture decoding device shown in FIG. To execute.

  FIG. 11 is a block diagram showing an example of an information processing apparatus that operates according to the second embodiment of the moving picture coding program of the present invention. In the figure, the same components as those in FIG. 9 are denoted by the same reference numerals, and the description thereof is omitted. The basic configuration of the information processing apparatus 900 illustrated in FIG. 11 is the same as that illustrated in FIG.

  However, in the central processing control device 903 shown in FIG. 11, the moving image coding program according to the second embodiment of the present invention is distributed from a recording medium or via a communication network and is taken in by the communication device 906. 9, motion estimation means 1101 corresponding to the motion estimator 502 shown in FIG. 5, motion compensation means 1102 equivalent to the motion compensator 503, and difference determination means equivalent to the difference determiner 504 Each of the functions 1103 is at least, and the same operation as that of the moving picture encoding apparatus shown in FIG. 5 is executed by software processing.

  FIG. 12 shows a block diagram of an example of an information processing apparatus that operates according to the second embodiment of the moving picture decoding program of the present invention. In the figure, the same components as those in FIG. 10 are denoted by the same reference numerals, and the description thereof is omitted. The basic configuration of the information processing apparatus 900 illustrated in FIG. 12 is the same as that illustrated in FIG.

  However, the central processing control apparatus 903 shown in FIG. 12 receives the moving picture decoding program according to the second embodiment of the present invention from a recording medium or via a communication network and is captured by the communication apparatus 906. 10 and at least the function of motion compensation means 1201 corresponding to the motion compensator 317 shown in FIG. 7, and the same operation as that of the moving picture decoding apparatus shown in FIG. Execute by processing.

  The present invention can be applied to a video encoding device, a video encoding program, and the like. Further, the present invention can be applied to a video decoding device, a video decoding program, and the like.

It is a block diagram of 1st Embodiment of the moving image encoder of this invention. It is a flowchart for operation | movement description of 1st Embodiment of the moving image encoder of this invention. It is a block diagram of 1st Embodiment of the moving image decoding apparatus of this invention. It is a flowchart for operation | movement description of 1st Embodiment of the moving image decoding apparatus of this invention. It is a block diagram of 2nd Embodiment of the moving image encoder of this invention. It is a flowchart for operation | movement description of 2nd Embodiment of the moving image encoder of this invention. It is a block diagram of 2nd Embodiment of the moving image decoding apparatus of this invention. It is a flowchart for operation | movement description of 2nd Embodiment of the moving image decoding apparatus of this invention. It is a block diagram of an example of the information processing apparatus which operate | moves by 1st Embodiment of the moving image encoding program of this invention. It is a block diagram of an example of the information processing apparatus which operate | moves by 1st Embodiment of the moving image decoding program of this invention. It is a block diagram of an example of the information processing apparatus which operate | moves by 2nd Embodiment of the moving image encoding program of this invention. It is a block diagram of an example of the information processing apparatus which operate | moves by 2nd Embodiment of the moving image decoding program of this invention. It is the conceptual diagram which showed the mode of the motion estimation which produces | generates motion vector information by performing the motion vector search by general block matching. It is a conceptual diagram for demonstrating the operation | movement of the area | region boundary motion estimation by the moving image encoder of this invention. It is a conceptual diagram for demonstrating an example which calculates the boundary condition in a boundary part at the time of performing area | region boundary motion estimation by the moving image encoder of this invention. It is a conceptual diagram for showing the operation | movement of the area | region boundary motion compensation by the moving image encoder of this invention. It is a conceptual diagram for demonstrating how to obtain | require the reference | standard difference information in the operation | movement of the difference motion estimation in the area | region by the moving image encoder of this invention. It is a conceptual diagram for showing the operation | movement in the difference motion estimation in the area | region by the moving image encoder of this invention. It is a conceptual diagram for showing another operation | movement in the difference motion estimation in the area | region by the moving image encoder of this invention. It is a conceptual diagram for demonstrating the operation | movement in the difference motion compensation in area | region by the moving image encoding apparatus of this invention, and the synthesis | combination of a prediction picture. It is a conceptual diagram for showing another operation | movement in the synthesis | combination of the difference motion compensation in a area | region by the moving image encoding apparatus of this invention, and a prediction picture. It is a block diagram of an example of the encoding apparatus at the time of introduce | transducing the smoothing process with respect to the prediction picture after motion compensation to the encoding apparatus of AVC. It is a flowchart for operation | movement description of an encoding apparatus at the time of introduce | transducing a smoothing process with respect to the prediction picture after a motion compensation to the encoding apparatus of AVC. It is a block diagram of an example of a decoding apparatus at the time of introduce | transducing smoothing processing with respect to the prediction picture after motion compensation to the decoding apparatus of AVC. It is a flowchart for operation | movement description of a decoding apparatus at the time of introduce | transducing the smoothing process with respect to the prediction picture after motion compensation to the decoding apparatus of AVC. It is a conceptual diagram for schematically representing a residual signal obtained from a difference between an estimated signal estimated from a Poisson equation by applying a boundary condition of a predetermined region and an original signal and an estimated signal. It is a conceptual diagram which represents typically the problem at the time of performing DCT with respect to an original signal. It is a conceptual diagram which represents typically the difference with the conventional method by performing DCT with respect to a residual signal. It is a conceptual diagram which represents typically the state at the time of dividing | segmenting each picture in a predetermined rectangular area.

Explanation of symbols

101, 301 Input code 102, 103, 302, 318, 501 Switch 104, 311 Frame memory 105, 308 Intra predictor 107 Orthogonal transformer 108 Quantizer 109, 306 Inverse quantizer 110, 307 Inverse orthogonal transformer 111, 309 First synthesizer 112, 310 Deblock filter 113 Entropy encoder 114 Multiplexer 115 Output transmitter 116, 312 Output code 117 Encoding controller 118 Region boundary motion estimator 119, 314 Region boundary motion compensator 120 1 subtractor 121 intra-regional differential motion estimator 122, 315 intra-regional differential motion compensator 123, 316 second combiner 124 second subtractor 303 input receiver 304 demultiplexer 305 entropy decoder 313 decoding controller 317, 503 motion compensator 02 motion estimator 504 difference determiner 900 information processing device 903 central processing control device 1301, 1401 coded picture 1302, 1402 reference picture 1303, 1403 prediction target 1304, 1404 reference position 1305, 1405 search range 1406 adapted boundary portion 1407 boundary Partial motion vector information 1601 to 1604 Boundary part in reference picture 1605 Predicted picture 1606 Predicted block 1702 Reference difference information 1801 Matched block 1802 Difference component motion vector information 1803 Current search target 1901 Current search within reference picture search range Regions that are spatial positions corresponding to the target block 2001, 2102 Prediction residual block 2002, 2103 Prediction residual picture 2003, 2104 Second prediction picture

Claims (8)

One screen area of the input video signal to be encoded is subdivided with a rectangular area having a predetermined number of pixels as a unit, and the rectangular area is generated from a reference picture that is a locally decoded image signal with a processing unit as a unit. In a moving picture coding apparatus that generates a difference picture that is a difference signal between a predicted picture and a coded picture that is a moving picture signal to be coded, and performs coding on the difference picture.
The boundary condition between the rectangular area to be predicted in the coded picture and the other rectangular area adjacent to the rectangular area is determined as a search unit, and the boundary condition is matched with the boundary condition A boundary part of the reference picture having a condition is specified by performing a motion vector search in the reference picture, and further from a boundary part of the rectangular area in the encoded picture to the specified boundary part in the reference picture Region boundary motion estimation means for generating boundary motion vector information that is motion vector information of
Each rectangular region in the encoded picture that specifies a boundary condition of a corresponding boundary part from the reference picture based on the boundary part motion vector information and satisfies the Poisson equation based on the boundary condition of the specified boundary part Region boundary motion compensation means for generating a first predicted picture by generating an estimated image signal of
Performing a difference operation between the signal component in the rectangular area in the encoded picture to be processed and the signal component in the rectangular area in the first predicted picture generated by the area boundary motion compensation means Generating reference difference information, which is difference information serving as a reference for comparison, a signal component in a rectangular area that is a motion vector search target in the reference picture, and a rectangle in the first predicted picture By performing the difference calculation with the signal component in the region, the search target difference information to be compared at the time of motion vector search is generated, and the search target difference information is matched with the signal component of the reference difference information. A motion vector search is performed within the reference picture for a rectangular area to be processed, and a differential component motion vector that is motion vector information up to the rectangular area within the adapted reference picture And area difference motion estimation means for generating a broadcast,
Based on the difference component motion vector information, a corresponding rectangular area is identified from the reference picture, a signal component in the identified rectangular area in the reference picture, and a corresponding rectangular area in the first predicted picture An intra-regional differential motion compensation means for generating a prediction residual picture that is a differential prediction picture from a difference from the signal component
Combining means for generating a second predicted picture by combining the first predicted picture and the prediction residual picture;
First difference means for generating a first difference picture from a difference between the second predicted picture and the encoded picture;
Orthogonal transform means for generating orthogonal transform coefficient information by performing predetermined orthogonal transform on the first difference picture;
Quantization means for generating post-quantization information by performing predetermined quantization on the orthogonal transform coefficient information based on a predetermined quantization parameter;
A predetermined inverse quantization is performed on the post-quantization information based on a predetermined quantization parameter to generate post-quantization information, and a predetermined inverse orthogonal transform is performed on the post-quantization information. Decoded differential picture generation means for generating a decoded differential picture;
Local decoded image signal generating means for combining the decoded differential picture and at least the second predicted picture to generate the locally decoded image signal;
The local decoded image signal for at least one picture is retained as the reference picture, and the retained reference picture is retained as the region boundary motion estimation unit, the region boundary motion compensation unit, the intra-regional difference motion estimation unit, and the Storage means for supplying each to the intra-region differential motion compensation means;
Entropy encoding means for generating an encoded bit string by performing predetermined entropy encoding on at least the post-quantization information, the boundary motion vector information, and the difference component motion vector information;
Multiplexing means for generating an encoded bitstream by multiplexing the encoded bit string based on a predetermined syntax structure;
A moving picture encoding apparatus comprising: Motion estimation means for performing motion vector search by block matching in the reference picture acquired from the storage means in each rectangular area in the encoded picture, and generating motion vector information up to a matching rectangular area;
Based on the motion vector information generated by the motion estimation unit, a corresponding rectangular region is identified from the reference picture acquired from the storage unit, and a motion compensation unit that generates a fourth prediction picture;
Second difference means for generating a second difference picture from the difference between the fourth predicted picture and the encoded picture;
A difference determination means for comparing the first difference picture and the second difference picture based on a predetermined determination criterion, and selecting a difference picture having a smaller amount of information;
Encoding control means for performing control for supplying a difference picture selected by the difference determination means and a prediction picture that is a generation element of the selected difference picture to a predetermined means;
The moving picture encoding apparatus according to claim 1, further comprising: One screen area of the input video signal to be encoded is subdivided with a rectangular area having a predetermined number of pixels as a unit, and the rectangular area is generated from a reference picture that is a locally decoded image signal with a processing unit as a unit. A moving picture code for causing a computer to generate a difference picture that is a difference signal between a prediction picture and a coded picture that is a moving picture signal to be coded, and to perform coding on the difference picture In the program
The computer,
The boundary condition between the rectangular area to be predicted in the coded picture and the other rectangular area adjacent to the rectangular area is determined as a search unit, and the boundary condition is matched with the boundary condition A boundary part of the reference picture having a condition is specified by performing a motion vector search in the reference picture, and further from a boundary part of the rectangular area in the encoded picture to the specified boundary part in the reference picture Region boundary motion estimation means for generating boundary motion vector information that is motion vector information of
Each rectangular region in the encoded picture that specifies a boundary condition of a corresponding boundary part from the reference picture based on the boundary part motion vector information and satisfies the Poisson equation based on the boundary condition of the specified boundary part Region boundary motion compensation means for generating a first predicted picture by generating an estimated image signal of
Performing a difference operation between the signal component in the rectangular area in the encoded picture to be processed and the signal component in the rectangular area in the first predicted picture generated by the area boundary motion compensation means Generating reference difference information, which is difference information serving as a reference for comparison, a signal component in a rectangular area that is a motion vector search target in the reference picture, and a rectangle in the first predicted picture By performing the difference calculation with the signal component in the region, the search target difference information to be compared at the time of motion vector search is generated, and the search target difference information is matched with the signal component of the reference difference information. A motion vector search is performed within the reference picture for a rectangular area to be processed, and a differential component motion vector that is motion vector information up to the rectangular area within the adapted reference picture And area difference motion estimation means for generating a broadcast,
Based on the difference component motion vector information, a corresponding rectangular area is identified from the reference picture, a signal component in the identified rectangular area in the reference picture, and a corresponding rectangular area in the first predicted picture An intra-regional differential motion compensation means for generating a prediction residual picture that is a differential prediction picture from the difference from the signal component of
Combining means for generating a second predicted picture by combining the first predicted picture and the prediction residual picture;
First difference means for generating a first difference picture from a difference between the second predicted picture and the encoded picture;
Orthogonal transform means for generating orthogonal transform coefficient information by performing predetermined orthogonal transform on the first difference picture;
Quantization means for generating post-quantization information by performing predetermined quantization on the orthogonal transform coefficient information based on a predetermined quantization parameter;
A predetermined inverse quantization is performed on the post-quantization information based on a predetermined quantization parameter to generate post-quantization information, and a predetermined inverse orthogonal transform is performed on the post-quantization information. Decoded differential picture generation means for generating a decoded differential picture;
Local decoded image signal generating means for combining the decoded differential picture and at least the second predicted picture to generate the locally decoded image signal;
Entropy encoding means for generating an encoded bit string by performing predetermined entropy encoding on at least the post-quantization information, the boundary motion vector information, and the difference component motion vector information;
In addition to functioning as multiplexing means for generating an encoded bitstream by multiplexing the encoded bit string based on a predetermined syntax structure, the storage device receives the locally decoded image signal for at least one picture as described above. Storage means for holding the reference picture as a reference picture and supplying the held reference picture to the region boundary motion estimation unit, the region boundary motion compensation unit, the intra-regional differential motion estimation unit, and the intra-regional differential motion compensation unit, respectively. A moving picture coding program which is made to function as: The computer,
Motion estimation means for performing motion vector search by block matching in the reference picture acquired from the storage means in each rectangular area in the encoded picture, and generating motion vector information up to a matching rectangular area;
Based on the motion vector information generated by the motion estimation unit, a corresponding rectangular region is identified from the reference picture acquired from the storage unit, and a motion compensation unit that generates a fourth prediction picture;
Second difference means for generating a second difference picture from the difference between the fourth predicted picture and the encoded picture;
A difference determination means for comparing the first difference picture and the second difference picture based on a predetermined determination criterion, and selecting a difference picture having a smaller amount of information;
The differential determination means further functions as an encoding control means for performing control for supplying a predetermined picture and a prediction picture that is a generation element of the selected difference picture to a predetermined means. Item 4. A moving image encoding program according to Item 3. An encoded bitstream generated by a moving image encoding apparatus according to claim 1 or 2 or a computer that executes the moving image encoding program according to claim 3 or 4 is transmitted from a predetermined storage medium or a predetermined transmission path. A video decoding device that acquires and outputs a decoded video signal by performing a decoding operation on the acquired encoded bitstream,
Demultiplexing means for demultiplexing encoded information based on a predetermined syntax structure for the encoded bitstream;
Predemultiplexed information obtained from the demultiplexing means is subjected to predetermined entropy decoding, and at least the post-quantization information, boundary motion vector information, difference component motion vector information, and predetermined syntax Entropy decoding means for generating parameter information necessary for constructing the structure;
A predetermined inverse quantization based on a predetermined quantization parameter is performed on the post-quantization information acquired from the entropy decoding unit to generate dequantized information, and the post-quantization information is predetermined. Decoding difference picture generation means for generating a decoding difference picture by performing inverse orthogonal transformation of
First synthesizing means for synthesizing the decoded differential picture and the synthesis prediction picture to generate a locally decoded image signal;
Storage means for holding the locally decoded image signal for at least one picture as a reference picture;
Based on the boundary part motion vector information acquired from the entropy decoding means, the boundary condition of the corresponding boundary part is specified from the reference picture, and the Poisson equation is satisfied based on the boundary condition of the specified boundary part A region boundary motion compensation unit that generates a first predicted picture by generating an estimated image signal in each rectangular region in an encoded picture that is a moving image signal to be encoded;
Based on the difference component motion vector information acquired from the entropy decoding means, a corresponding rectangular area is identified from the reference picture, a signal component in the identified rectangular area, and the first predicted picture An intra-regional differential motion compensation means for generating a prediction residual picture that is a differential prediction picture from a difference from a signal component in a corresponding rectangular region in
Second synthesizing means for synthesizing the first predicted picture and the prediction residual picture, thereby generating a second predicted picture to be supplied to the first synthesizing means as the synthesis predicted picture;
A moving picture decoding apparatus comprising: The entropy decoding means further has a function of decoding the motion vector information,
Motion compensation means for identifying a corresponding rectangular area from the reference picture based on the motion vector information acquired from the entropy decoding means, and generating a fourth predicted picture;
Decoding control information necessary for decoding control is acquired from parameter information necessary for constructing a syntax structure decoded by the entropy decoding unit, and supplied to the first combining unit according to the decoding control information Decoding control means for selecting one of the second predicted picture and the fourth predicted picture as the synthesis predicted picture.
The moving picture decoding apparatus according to claim 5, further comprising: An encoded bitstream generated by a moving image encoding apparatus according to claim 1 or 2 or a computer that executes the moving image encoding program according to claim 3 or 4 is transmitted from a predetermined storage medium or a predetermined transmission path. A moving picture decoding program for causing a computer to perform a decoding operation on the acquired encoded bitstream and outputting a decoded moving picture signal,
The computer,
Demultiplexing means for demultiplexing encoded information based on a predetermined syntax structure for the encoded bitstream;
Predemultiplexed information obtained from the demultiplexing means is subjected to predetermined entropy decoding, and at least the post-quantization information, boundary motion vector information, difference component motion vector information, and predetermined syntax Entropy decoding means for generating parameter information necessary for constructing the structure;
A predetermined inverse quantization based on a predetermined quantization parameter is performed on the post-quantization information acquired from the entropy decoding unit to generate dequantized information, and the post-quantization information is predetermined. Decoding difference picture generation means for generating a decoding difference picture by performing inverse orthogonal transformation of
First synthesizing means for synthesizing the decoded differential picture and at least the synthesis prediction picture to generate a locally decoded image signal;
Based on the boundary part motion vector information acquired from the entropy decoding means, the boundary condition of the corresponding boundary part is specified from the reference picture, and the Poisson equation is satisfied based on the boundary condition of the specified boundary part A region boundary motion compensation unit that generates a first predicted picture by generating an estimated image signal in each rectangular region in an encoded picture that is a moving image signal to be encoded;
Based on the difference component motion vector information acquired from the entropy decoding means, a corresponding rectangular area is identified from the reference picture, a signal component in the identified rectangular area, and the first predicted picture An intra-regional differential motion compensation means for generating a prediction residual picture that is a differential prediction picture from a difference from a signal component in a corresponding rectangular region in
Second synthesizing means for synthesizing the first predicted picture and the prediction residual picture, thereby generating a second predicted picture to be supplied to the first synthesizing means as the synthesis predicted picture;
A moving picture decoding program characterized by causing the storage device to function as storage means for holding the local decoded image signal for at least one picture as a reference picture. The computer,
Further generating the motion vector information by the entropy decoding means,
Motion compensation means for identifying a corresponding rectangular area from the reference picture based on the motion vector information acquired from the entropy decoding means, and generating a fourth predicted picture;
Decoding control information necessary for decoding control is acquired from parameter information necessary for constructing a syntax structure decoded by the entropy decoding unit, and supplied to the first combining unit according to the decoding control information Decoding control means for selecting one of the second predicted picture and the fourth predicted picture as the synthesis predicted picture.
8. The moving picture decoding program according to claim 7, further causing the function to function.

Images