抄録
This paper reports the results of a study on improvement of disparity estimation using adaptive processing. When the high-pass filtered luminance signals are used in block matching based on the sum of absolute error criteria, the accuracy of disparity estimation can be improved around edges in the images. In order to reduce miss-matching, we propose an adaptive processing algorithm to determine the proper function for each block. We verified the effect of this algorythm on the disparity-compensated images.
