| 摘要 |
The present invention includes the following steps: a step of generating a first motion vector candidate MVC1 based on a reference frame Rf1 and a second motion vector candidate MVC2 based on a second reference frame Rf2 (S300); a step of generating an interpolation predictive block PB0 by performing interpolation on each pixel value of each pixel in a block PB1 for prediction, indicated by the first motion vector candidate MVC1 and each pixel value of each pixel in a block PB2 for prediction indicated by the second motion vector candidate MVC2 (S302, S304); a step of calculating a prediction error evaluation value based on a differential between each pixel value in the interpolation predictive block PB0 and each pixel value in a current block to be coded (S306 SIMILAR S310); and a step of estimating, for the current block, the first motion vector candidate MVC1 and the second motion vector candidate MVC2, based on which the smallest prediction error evaluation value is obtained, respectively as a motion vector MV1 and a motion vector MV2 (S312, S314). <IMAGE> |
