Segment-Based Stereo Matching Using Belief Propagation and a Self-Adapting Dissimilarity Measure

TLDR

The paper proposes a stereo matching algorithm that uses color segmentation and a self‑adapted matching score to maximize reliable correspondences. The method models the scene as planar surface patches, assigns a disparity plane to each segment, estimates the patches robustly to outliers, and optimizes the labeling with belief propagation. Experimental results on the Middlebury stereo test bed demonstrate the superior performance of the proposed method.

Abstract

A novel stereo matching algorithm is proposed that utilizes color segmentation on the reference image and a self-adapting matching score that maximizes the number of reliable correspondences. The scene structure is modeled by a set of planar surface patches which are estimated using a new technique that is more robust to outliers. Instead of assigning a disparity value to each pixel, a disparity plane is assigned to each segment. The optimal disparity plane labeling is approximated by applying belief propagation. Experimental results using the Middlebury stereo test bed demonstrate the superior performance of the proposed method