| 摘要 |
The present invention relates to an image processing apparatus and method in which an increase in the amount of encoding can be suppressed while realizing high-performance motion compensation. In step S1, a motion vector corresponding to the difference between the coordinates of an encoding target block and the coordinates of a reference block is calculated. In step S2, a horizontal-direction fine adjustment filter Hh is generated. In step S3, a vertical-direction fine adjustment filter Hv is generated. In step S4, the pixel values of the reference block is finely adjusted by applying the horizontal-direction and vertical-direction fine adjustment filters Hh and Hv to the reference block. The reference block is then assigned to the encoding target block, thereby generating a motion-compensated prediction image. In step S5, the difference signal between the pixel values of the encoding target image and the pixel values of the prediction image is calculated, and is subjected to orthogonal transform and quantization. The resulting difference signal is then encoded. Meanwhile, the integer-precision motion vector is also encoded. The present invention is applicable to an image encoding apparatus for compression-encoding moving images and to an image decoding apparatus for decoding the encoded moving images.
|
