Image coding apparatus, image coding method, image decoding apparatus, image decoding method and communication apparatus

Abstract

Each of an image coding apparatus and an image decoding apparatus uses a motion compensated prediction using virtual samples so as to detect a motion vector for each of regions of each frame of an input signal. Accuracy of virtual samples is locally determined while the accuracy of virtual samples is associated with the size of each region which is a motion vector detection unit in which a motion vector is detected. Virtual samples having half-pixel accuracy are used for motion vector detection unit regions having a smaller size 8×8 MC, such as blocks of 8×4 size, blocks of 4×8 size, and blocks of 4×4 size, and virtual samples having ¼-pixel accuracy are used for motion vector detection unit regions that are equal to or larger than 8×8 MC in size.

1 . A moving image coding apparatus that performs a motion compensated prediction on each of regions, into which each frame of a moving picture signal is divided according to a predetermined method, so as to generate a predicted picture, and that compresses and encodes a differential signal indicating a difference between the generated, predicted picture and said moving picture signal so as to generate a coded bitstream, said moving image coding apparatus comprising: 
 a frame memory for storing a reference picture used for generating the predicted picture;    a motion detecting unit for changing accuracy of virtual pixels that becomes elements, which constitute the predicted picture, according to a shape of each of the regions that are units for the motion compensated prediction, so as to generate a plurality of candidates for the predicted picture, and for generating a motion vector that provides one of said plurality of candidates for the predicted picture which has a higher degree of prediction efficiency; and    a motion compensating unit for changing the accuracy of virtual pixels that become elements, which constitute the predicted picture, according to the shape of each of the regions that are units for the motion compensated prediction, so as to generate the predicted picture based on the motion vector generated by said motion detecting unit, said apparatus multiplexing shape information indicating the shape of each of the regions that are units for the motion compensated prediction and said motion vector into said coded bitstream.    
   
   
       2 . The moving image coding apparatus according to  claim 1 , wherein said apparatus multiplexes the shape information indicating the shape of each of the regions that are units for the motion compensated prediction and said motion vector that is coded by using a predictively coding method that is changed according to the shape of each of the regions that are units for the motion compensated prediction, which is indicated by the shape information, into said coded bitstream.  
   
   
       3 . The moving image coding apparatus according to  claim 1 , wherein said motion detecting unit generates the plurality of candidates for the predicted picture based on a control signal used for determining whether to change the accuracy of virtual pixels that become elements, which constitute the predicted picture, according to the shape of each of the regions that are units for the motion compensated prediction and in units of predetermined moving picture data, and generates the motion vector that provides one of said plurality of candidates for the predicted picture which has a higher degree of prediction efficiency, and said motion compensating unit generates the predicted picture based on the motion vector generated by said motion detecting unit and based on the control signal used for determining whether to change the accuracy of virtual pixels that become elements, which constitute the predicted picture, according to the shape of each of the regions that are units for the motion compensated prediction and in units of predetermined moving picture data.  
   
   
       4 . The moving image coding apparatus according to  claim 1 , wherein said motion compensating unit performs motion compensated predictions on a region-by-region basis according to a specified one of a first degree of accuracy with which said motion compensating unit generates virtual pixels based on plural pixel data about the reference picture stored in said frame memory by using a predetermined method, and a second degree of accuracy with which said motion compensating unit generates virtual pixels based on the virtual pixels generated with the first degree of accuracy, so as to generate the reference picture.  
   
   
       5 . The moving image coding apparatus according to  claim 1 , wherein each of the regions that are units for the motion compensated prediction is one of a plurality of prediction unit blocks into which a macroblock that is a 16 pixel×16 line region of a luminance component of the moving picture signal of each frame is further divided, and the shape information indicating the shape of each of the regions is information used for specifying a method of dividing the macroblock into the plurality of prediction unit blocks.  
   
   
       6 . The moving image coding apparatus according to  claim 1 , wherein said apparatus includes a plurality of frame memories each for storing a reference picture used for generating the predicted picture, and said motion compensating unit performs the motion compensated prediction by referring to a plurality of reference pictures stored in said plurality of frame memories so as to generates the predicted picture.  
   
   
       7 . The moving image coding apparatus according to  claim 1 , wherein said apparatus has an intra mode in which said apparatus intra-codes the moving picture signal, and selects either a motion compensated prediction mode using said motion compensating unit or said intra mode and encodes the moving picture signal in the selected motion compensated prediction mode or intra mode, and further multiplexes coding mode information indicating the selected mode into said coded bitstream.  
   
   
       8 . The moving image coding apparatus according to  claim 1 , wherein said apparatus further includes a spatial prediction unit for predictively coding the moving picture signal in a spatial predictionmode, and selects either amotion compensated prediction mode using said motion compensating unit or the spatial prediction mode using said spatial prediction unit and encodes the moving picture signal in the selected motion compensated prediction mode or spatial prediction mode, and further multiplexes coding mode information indicating the selected mode into said coded bitstream.  
   
   
       9 . A moving picture decoding apparatus that accepts a coded bitstream which is generated from a compressed and coded differential signal indicating a difference between a predicted picture and a moving picture signal, the predicted picture being generated by using a motion compensated prediction for each of regions, into which each trame of said moving picture signal is divided according to a predetermined method, so as to reconstruct the moving picture signal, said moving picture decoding apparatus comprising: 
 a frame memory for storing a reference picture used for generating the predicted picture;    a decoding unit for accepting and decoding said coded bitstream so as to reconstruct said differential signal, a motion vector, and shape information indicating a shape of each of the regions that are units for the motion compensated prediction; and    a motion compensating unit for changing accuracy of virtual pixels that become elements, which constitute the predicted picture, according to the shape of each of the regions that are units for the motion compensated prediction, and for generating the predicted picture according to the changed accuracy by using the motion vector reconstructed by said decoding unit and by referring to the reference picture stored in said frame memory, said apparatus reconstructing the moving picture signal by adding the predicted picture generated by said motion compensating unit to said differential signal reconstructed by said decoding unit.    
   
   
       10 . The moving picture decoding apparatus according to  claim 9 , wherein said decoding unit accepts and decodes the coded bitstream so as to reconstruct the differential signal and the shape information indicating the shape of each of the regions that are units for the motion compensated prediction, and also reconstruct the motion vector by changing a predictively reconstructing method of predictively reconstructing the motion vector based on said shape information.  
   
   
       11 . The moving picture decoding apparatus according to  claim 9 , wherein said decoding unit accepts and decodes the coded bitstream so as to reconstruct the differential signal, the shape information indicating the shape of each of the regions that are units for the motion compensated prediction, and a control signal that is defined in units of predetermined moving picture data, determines whether to change a predictively reconstructing method of predictively reconstructing the motion vector according to said shape information, based on said control signal and in units of predetermined moving picture data so as to reconstruct the motion vector, and wherein said motion compensating unit determines whether to change the accuracy of virtual pixels that become elements, which constitute the predicted picture, according to the shape of each of the regions that are units for the motion compensated prediction, based on said control signal and in units of predetermined moving picture data so as to determine the accuracy of virtual pixels, and generates the predicted picture by using the motion vector, which is reconstructed according to the determined accuracy by said decoding unit, and by referring to the reference picture stored in said frame memory.  
   
   
       12 . The moving picture decoding apparatus according to  claim 9 , wherein said motion compensating unit performs a motion compensation on a region-by-region basis according to a specified one of a first degree of accuracy with which said motion compensating unit generates virtual pixels based on plural pixel data about the reference picture stored in said frame memory by using a predetermined method, and a second degree of accuracy with which said motion compensating unit generates virtual pixels based on the virtual pixels generated with the first degree of accuracy, so as to generate the reference picture.  
   
   
       13 . The moving image coding apparatus according to  claim 9 , wherein each of the regions that are units for the motion compensated prediction is one of a plurality of prediction unit blocks into which a macroblock that is a 16 pixel×16 line region of a luminance component of each frame of the moving picture signal is further divided, the shape information indicating the shape of each of the regions is information used for specifying a method of dividing the macroblock into the plurality of prediction unit blocks, and the corresponding motion vector is a motion vector used in each of the plurality of prediction unit blocks.  
   
   
       14 . The moving picture decoding apparatus according to  claim 9 , wherein said apparatus includes a plurality of frame memories each for storing a reference picture used for generating the predicted picture, and said motion compensating unit performs a motion compensation by referring to a plurality of reference pictures stored in said plurality of frame memories so as to generate the predicted picture.  
   
   
       15 . The moving picture decoding apparatus according to  claim 9 , wherein said decoding unit further reconstructs coding mode information from said coded bitstream and then reconstructs the moving picture signal in either an intra mode or a motion compensated prediction mode using said motion compensating unit according to said coding mode information.  
   
   
       16 . The moving picture decoding apparatus according to  claim 9 , wherein said apparatus further includes a spatial prediction unit for predictively coding the moving picture signal in a spatial prediction mode, and reconstructs coding mode information from said coded bitstream and then reconstructs the moving picture signal in either the spatial prediction mode using said spatial prediction unit or a motion compensated prediction mode using said motion compensating unit according to said coding mode information.  
   
   
       17 . A moving picture decoding method of accepting a coded bitstream which is generated from a compressed and coded differential signal indicating a difference between a predicted picture and a moving picture signal, the predicted picture being generated by using a motion compensated prediction for each of regions, into which each frame of said moving picture signal is divided according to a predetermined method, so as to reconstruct the moving picture signal, said moving picture decoding method comprising the steps of: 
 accepting and decoding said coded bitstream so as to reconstruct said differential signal, a motion vector, and shape information indicating a shape of each of the regions that are units for the motion compensated prediction;    changing accuracy of virtual pixels that become elements, which constitute the predicted picture, according to the shape of each of the regions that are units for the motion compensated prediction; generating the predicted picture according to the changed accuracy by using the reconstructed motion vector and by referring to a reference picture; and    reconstructing the moving picture signal by adding the predicted picture generated by using the motion compensation to said reconstructed differential signal.    
   
   
       18 . A communication apparatus comprising the image coding apparatus according to  claim 1 .  
   
   
       19 . A communication apparatus comprising the image decoding apparatus according to  claim 9.