| 摘要 |
In a method for predictive video compression coding, image information is transmitted as successive image blocks each comprising n x m pixels, and (i) previously transmitted image information is searched to find a prediction block for a pixel block (A) to be transmitted by comparing the block with at least one previously transmitted image area (B) of equal size, and (ii) information proportional to the differences between the intensity values of the corresponding pixels of the block (A) to be transmitted and the found prediction block is transmitted to a receiver. In order that blocks corresponding in pattern to the block to be coded but differing from it in basic illumination could be used as a prediction block, and in order that the accuracy of the prediction blocks could be improved, the comparison is performed on the basis of pixel intensity values obtained by subtracting from the original intensity values (Ai,j, Bi,j) a value representing the basic illumination of each image block, and that when the prediction block is found, (i) a difference image calculated in a manner known per se pixel by pixel on the basis of the original pixel intensity values (Ai,j, Bi,j) is transmitted to the receiver, or (ii) a difference image calculated in a manner known per se pixel by pixel on the basis of said pixel intensity values is transmitted to the receiver as well as a difference value (DCe) of the values representing the basic illuminations of the block to be transmitted and the prediction block. <IMAGE> |
