摘要:
A learning sample collector is configured to collect learning samples which have been classified into respective classes through N-classes discrimination (N is a natural number of 3 or more). A classifier is configured to reclassify the learning samples collected by the learning sample collector into classes applied to M-classes discrimination, where M is smaller than N (M is a natural number of 2 or more and is less than N). A learner is configured to learn a discriminator for performing the M-classes discrimination on a basis of the learning samples reclassified by the classifier.
摘要:
When an intra prediction parameter indicates a horizontal prediction process, an intra prediction unit adds a value proportional to a change in a horizontal direction of the luminance values of pixels adjacent to the top of each block, which is a unit for prediction process of a coding block, to the luminance value of a pixel adjacent to the left of the block to set a result of the addition as a predicted value of a prediction image, whereas when the intra prediction parameter indicates a vertical prediction process, the intra prediction unit adds a value proportional to a change in a vertical direction of the luminance values of pixels adjacent to the left of the block to the luminance value of the pixel adjacent to the top of the block to set the result of the addition as a predicted value of the prediction image.
摘要:
An encoding controlling unit 3 selects one transformation block size which provides an optimal degree of encoding efficiency from a set of transformation block sizes which are determined in accordance with an encoding mode 7, and includes the transformation block size selected thereby in optimal compression parameters 20a to notify the transformation block size to a transformation/quantization unit 19, and the transformation/quantization unit 19 divides an optimal prediction differential signal 13a into blocks having the transformation block size included in the optimal compression parameters 20a, and carries out a transformation and quantization process on each of the blocks to generate compressed data 21.
摘要:
When a randomly-accessible inter picture is encoded, a prediction image generator sets a reference picture from among a plurality of randomly-accessible intra pictures, and performs a motion-compensated prediction using the set reference picture for a prediction process, and a variable length encoding unit 13 encodes both picture position information showing the position of the reference picture, and identification information showing that the randomly-accessible inter picture is randomly accessible, and multiplexes encoded data about both the picture position information and the identification information into a bitstream.
摘要:
When carrying out an average prediction, an intra predictor carries out a filtering process on target pixels of the intra prediction located at an upper end and a left end of the block, the filtering process using an intermediate prediction value, which is an average value of adjacent pixels of the block, and at least one adjacent pixel of the target pixel. The intra predictor also sets a filter coefficient to ¾, associated with the intermediate prediction value for a target pixel at the upper end of the block other than the target pixel at an upper left corner of the block, and sets a filter coefficient to ¼, associated with the adjacent pixel adjacent to the upper side of the target pixel at the upper end of the block. As a result, prediction errors locally occurring can be reduced, and the image quality can be improved.
摘要:
When an encoding mode corresponding to one of blocks to be encoded into which an image is divided by a block dividing part 2 is an inter encoding mode which is a direct mode, a motion-compensated prediction part 5 selects a motion vector suitable for generation of a prediction image from one or more selectable motion vectors and also carries out a motion-compensated prediction process on the block to be encoded to generate a prediction image by using the motion vector, and outputs index information showing the motion vector to a variable length encoding part 13, and the variable length encoding unit 13 variable-length-encodes the index information.
摘要:
An image encoding device include a predicting unit for adaptively determining the size of each motion prediction unit block according to color component signals, and for dividing each motion prediction unit block into motion vector allocation regions to search for a motion vector, and a variable length encoding unit for, when a motion vector is allocated to the whole of each motion prediction unit block, performing encoding in mc_skip mode if the motion vector is equal to an estimated vector and a prediction error signal does not exist, and for, when each motion vector allocation region has a size equal to or larger than a predetermined size and a motion vector is allocated to the whole of each motion vector allocation region, performing encoding in sub_mc_skip mode if the motion vector is equal to an estimated vector and a prediction error signal does not exist.
摘要:
An image encoding device include a predicting unit for adaptively determining the size of each motion prediction unit block according to color component signals, and for dividing each motion prediction unit block into motion vector allocation regions to search for a motion vector, and a variable length encoding unit for, when a motion vector is allocated to the whole of each motion prediction unit block, performing encoding in mc_skip mode if the motion vector is equal to an estimated vector and a prediction error signal does not exist, and for, when each motion vector allocation region has a size equal to or larger than a predetermined size and a motion vector is allocated to the whole of each motion vector allocation region, performing encoding in sub_mc_skip mode if the motion vector is equal to an estimated vector and a prediction error signal does not exist.
摘要:
An apparatus to decode an encoded bit stream obtained by dividing the moving image into a plurality of rectangular regions includes a rectangular region size determination unit for determining a variable size of the regions in a sequence (a set of continuous frames) based on region size information multiplexed in the encoded bit stream. The apparatus includes a decoding unit for decoding coding mode information multiplexed in the encoded bit stream for each of the regions, and a decoding unit for decoding pixel values in each of the regions based on the coding mode information. The coding mode information includes (1) prediction type information selected from information indicating either inter or intra prediction coding, and (2) intra prediction block size information indicating a size of a unit region for the intra prediction coding indicated by the prediction type information.
摘要:
A parallel processing starting unit 3 that partitions an inputted image into tiles each having a predetermined size, and distributes tiles obtained through the partitioning, and N tile encoding units 5-1 to 5-N each of that carries out a prediction difference encoding process on a tile distributed thereto by the parallel processing starting unit 3 to generate a local decoded image are disposed, and each of N tile loop filter units 7-1 to 7-N determines a filter per tile suitable for a filtering process on the local decoded image generated by the corresponding one of the tile encoding units 5-1 to 5-N, and carries out the filtering process on the local decoded image by using the filter.