Gradient Magnitude Differences and Guided Filter for Stereo Video Matching Algorithm

Abstract

This paper proposes a new stereo video matching algorithm which uses Gradient Magnitude (GM) differences and Guided Filter (GF). The radiometric and edges distortions are the problems that contribute to the quality of the results for stereo video matching algorithm. Hence, this article proposes an algorithm to reduce these problems. The first stage, the GM is utilized. The GM is strong against the radiometric distortion on an image due to different brightness on an image or between the stereo cameras. The second stage, the GF is used to improve the edges of object matching and is efficiently to remove the noise. Based on the standard benchmarking stereo dataset, the proposed work in this article produces good results and performs much better compared to before the proposed framework. This new algorithm is also competitive with some established methods in the literature.