Image encoding device, image decoding device, image encoding method, and image decoding method

An image encoding device comprising:

• a prediction image generating unit having a first prediction image generating method of searching for a motion vector between a block to be encoded and a reference image for each block which constructs an inputted image, and carrying out a motion-compensated prediction process using said motion vector to generate a prediction image, and a second prediction image generating method of generating a direct vector of the block to be encoded from a motion vector of an already-encoded block, and carrying out a motion-compensated prediction process using said direct vector to generate a prediction image, said prediction image generating unit generating a prediction image only by using said second prediction image generating method without generating a prediction image by using said first prediction image generating method when control information for instructing a use of only said second prediction image generating method is inputted;
• a difference image selecting unit for evaluating a degree of efficiency of encoding of a difference image which is a difference between an image of the block to be encoded and the prediction image generated by said prediction image generating unit by using the first prediction image generating method, and also evaluating a degree of efficiency of encoding of a difference image which is a difference between the image of the block to be encoded and the prediction image generated by said prediction image generating unit by using the second prediction image generating method to select a difference image having a higher one of said degrees of encoding efficiency;
• a quantizing unit for quantizing the difference image selected by said difference image selecting unit to output quantization coefficients of said difference image;
• and
• a variable length encoding unit for variable-length-encoding prediction image identification information showing whether the prediction image associated with the quantization coefficients outputted from said quantizing unit and the difference image selected by said difference image selecting unit is either the one generated by using the first prediction image generating method or the one generated by using the second prediction image generating method, and, when the difference image selected by said difference image selecting unit is the one associated with the prediction image generated by using the first prediction image generating method, variable-length-encoding the motion vector which is searched for by said prediction image generating unit to generate encoded data.

An image decoding device which receives, as an input, a bitstream which is acquired by compression-encoding each predetermined block of an image by using a motion-compensated prediction, and which carries out a decoding expansion process on said each block, said image decoding device comprising:

• a first prediction image generating unit for decoding a motion vector of said each block from the bitstream, and for generating a first prediction image by using said motion vector;
• a decoding unit for decoding encoded data about a difference image for motion-compensated prediction of said each block from the bitstream, and for generating a decoded difference image from said encoded data;
• an adding unit for adding said first prediction image and said decoded difference image to acquire a decoded image;
• a memory for holding a motion vector which is acquired by decoding said bitstream;
• a second prediction image generating unit for generating a second prediction image by using a motion vector of a block to be decoded which is generated on a basis of the motion vector held by said memory;
• and
• a filtering unit for carrying out a smoothing process on a block boundary in the second prediction image generated by said second prediction image generating unit, wherein
• for each predetermined image area which consists of one or more of said blocks, said image decoding device decodes identification information showing that either the decoded image outputted from said adding unit or an image outputted from said filtering unit is a final decoded image in all of said one or more blocks in said image area from the bitstream, and generates a final decoded image on a basis of said identification information.

An image decoding device comprising:

• a variable length decoding unit for decoding quantization coefficients and prediction image identification information showing whether an image encoding device generates a prediction image by using either a first prediction image generating method or a second prediction image generating method from encoded data, and for, when the prediction image is generated by using the first prediction image generating method, decoding a motion vector from the encoded data;
• an inverse quantizing unit for inverse-quantizing the quantization coefficients decoded by said variable length decoding unit;
• a prediction image generating unit for carrying out a motion-compensated prediction process using the motion vector decoded by said variable length decoding unit to generate a prediction image when the prediction image identification information decoded by said variable length decoding unit shows that the image encoding device generates the prediction image by using the first prediction image generating method, and for generating a direct vector of a block to be decoded from a motion vector of an already-decoded block and carrying out a motion-compensated prediction process using said direct vector to generate a prediction image when the prediction image identification information decoded by said variable length decoding unit shows that the image encoding device generates the prediction image by using the second prediction image generating method;
• and
• a decoded image generating unit for handling the prediction image generated by the prediction image generating unit as a decoded image when the prediction image identification information decoded by said variable length decoding unit shows that said image encoding device generates the prediction image by using the second prediction image generating method.

An image encoding method comprising:

• a first prediction image generating process step of a prediction image generating unit implementing a first prediction image generating method of searching for a motion vector between a block to be encoded and a reference image for each block which constructs an inputted image, and carrying out a motion-compensated prediction process using said motion vector to generate a prediction image;
• a second prediction image generating process step of said prediction image generating unit implementing a second prediction image generating method of generating a direct vector of the block to be encoded from a motion vector of an already-encoded block, and carrying out a motion-compensated prediction process using said direct vector to generate a prediction image;
• a difference image selecting process step of a difference image selecting unit evaluating a degree of efficiency of encoding of a difference image which is a difference between an image of the block to be encoded and the prediction image generated in said first prediction image generating process step, and also evaluating a degree of efficiency of encoding of a difference image which is a difference between the image of the block to be encoded and the prediction image generated in said second prediction image generating process step to select a difference image having a higher one of said degrees of encoding efficiency;
• a quantizing process step of a quantizing unit quantizing the difference image selected in said difference image selecting process step to output quantization coefficients of said difference image;
• and
• a variable length encoding process step of a variable length encoding unit variable-length-coding prediction image identification information showing whether or not the prediction image associated with the quantization coefficients outputted in said quantizing process step and the difference image selected in said difference image selecting process step is the one generated by using the second prediction image generating method, and, when the difference image selected by said difference image selecting unit is the one associated with the prediction image generated by using the first prediction image generating method, variable-length-encoding the motion vector which is searched for in said prediction image generating process step to generate encoded data.

An image decoding device comprising:

• a variable length decoding process step of a variable length decoding unit decoding quantization coefficients and prediction image identification information showing whether an image encoding device generates a prediction image by using either a first prediction image generating method or a second prediction image generating method from encoded data, and, when the prediction image is generated by using the first prediction image generating method, decoding a motion vector from the encoded data;
• an inverse quantizing process step of an inverse quantizing unit inverse-quantizing the quantization coefficients decoded in said variable length decoding process step;
• a prediction image generating process step of a prediction image generating unit generating a direct vector of a block to be decoded from a motion vector of an already-decoded block and carrying out a motion-compensated prediction process using said direct vector to generate a prediction image when the prediction image identification information decoded in said variable length decoding process step shows that the image encoding device generates the prediction image by using the second prediction image generating method;
• and
• a decoded image generating process step of a decoded image generating unit adding the prediction image generated in said prediction image generating process step and a difference image shown by a result of the inverse quantization carried out in said inverse quantizing process step to generate a decoded image corresponding to an inputted image inputted to said image encoding device.