A Novel Trilateral Filter based Adaptive Support Weight Method for Stereo Matching

Abstract

Adaptive support weight (ASW) approach represents the state-of-the-art local stereo matching method. Recent extensive evaluation studies on ASW approaches show that the bilateral filter weight function enables outstanding performance on a large dataset in comparison with various weight functions. However, it does not resolve the ambiguity induced by nearby pixels at different disparities but with similar colors. In this paper, we propose a novel trilateral filter based ASW method which remedies such ambiguities by considering disparity discontinuities through color discontinuity boundaries, i.e., the strength of the boundary between two pixels. The experimental evaluation on the Middlebury benchmark shows that the proposed algorithm ranks 15th out of 150 submissions and is the current most accurate local stereo matching algorithm.