Disparity Refinement Through Grouping Areas and Support Weighted Windows

Abstract

In this work, we propose a simple but an effective technique to adjust a disparity map in a more appropriate configuration. This proposal consists of three main steps: segmentation process, statistical analysis and by using adaptive weighted windows. Furthermore, we investigate if a disparity map, yielded by a robust stereo method, can be improved by the proposed methodology. Thus, we implement some stereo vision methods to compare. The experimental results show that the proposed method is efficient and it can make some enhancements in disparity maps, as reducing the disparity error measure.