| 摘要 |
An image decoding apparatus and method are capable of reducing block noise without difference in gradation being lost, wherein, an image is divided into a plurality of blocks containing a plurality of pixels and the pixels are encoded with label information corresponding to representative values which represent gradation of pixels in the block, and when the size in the dynamic range (distribution range of gradation) of a focused block is not less than a prescribed multiple times the size in the dynamic range in an adjacent block, a representative value, in the focused block, whose difference from the intermediate value of representative values of the adjacent block is not more than a threshold value, is assumed to be adjusted, and a variance is provided to the representative value in a decoding process.
|
