摘要: In this paper we present a new stereo matching algorithm that produces accurate dense disparity maps and explicitly detects occluded areas. This extends the original cooperative algorithms in two ways. First, design method of adjusting initial score volume to guarantee correct matches have high scores. propagates "good" information within or among image segments based on certain confidence measurement criterion, thus improving robustness algorithm. Second, develop scheme choosing local support areas by enforcing segmentation information. sees depth discontinuities coincide with color intensity boundaries. As result, foreground fattening errors are drastically reduced. Extensive experimental results demonstrate effectiveness our algorithm, both quantitatively qualitatively. Comparison between some other representative is also reported.