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

A moving picture encoder generating an encoded bitstream by encoding a differential signal between a predicted picture and a frame of a moving picture signal, the encoder comprising:

• a frame memory storing a plurality of reference pictures available to generation of the predicted picture;
• a motion compensating unit detecting a motion vector between a frame to be encoded and the reference pictures stored in the frame memory, and generating the predicted picture by a motion compensated prediction on a region-by-region basis, each region being obtained by dividing each frame of the moving picture signal by a predetermined method;
• and
• an encoding unit transforming and quantizing the differential signal between the frame to be encoded and the predicted picture generated by the motion compensating unit, and generating variable-length encoded data of the differential signal, and multiplexing encoded parameters containing the motion vector and the variable-length encoded data into the bitstream,
• wherein the motion compensating unit including a mechanism selectively carrying out switching, on a frame-by-frame basis, to both fractional pixel accuracy of a fractional pixel and an interpolation filtering method to be used for generation of the predicted picture containing the fractional pixels, and switching, on a frame-by-frame basis, to the fractional pixel accuracy of the fractional pixel without switching the interpolation filtering method based on shape and size of each region without mirroring pixels placed near boundaries of each region, the mechanism detecting the motion vector based on both the fractional pixel accuracy and the interpolation filtering method defined on the frame-by-frame basis with reference to the reference pictures stored in the frame memory, and generating the predicted pictures each corresponding to respective reference pictures, and
• wherein the encoding unit multiplexes information indicating both the switched fractional pixel accuracy and the interpolation filtering method selected by the mechanism into the bitstream, encodes the detected motion vector of each region based on the information indicating both the switched fractional pixel accuracy and the interpolation filtering method selected by the mechanism, and encodes the information with the highest degree of macroblock encode efficiency from among a plurality of encoding modes indicating which of the reference pictures is used for generation of the predicted picture, and multiplexes the detected motion vector and the encoded information with the highest degree of macroblock encode efficiency into the bitstream.