Methods to improve stereo matching algorithm for object segmentation

摘要

Up to these days, there are huge amount of stereo matching algorithms and its improved versions. Not sufficient algorithms are proposed to be used in real systems like object recognition not considering time computing and complexity of the algorithm, but only for improving Pixel Error Rate (PER) in disparity image. We used some image processing methods for not only improving PER but also making disparity images good to be used for object segmentation. We used Trellis algorithm which is Dynamic Programming based stereo matching algorithm. Our proposed algorithm shows better results than original Trellis algorithm in qualitative and quantitative manner. Also this algorithm could be implemented in hardware system; it could work in real-time system.