Stereo matching using sift and rotation invariant uniform LTP

Abstract

This paper presents a novel method for stereo matching, We introduce the extension version of the well-known local binary pattern (LBP) feature called local ternary pattern (LTP) serving as s distinctive local descriptor of local image feature point region detected by the scale invariant feature transformation (SIFT) detector. In our experiments, our descriptor combined with the advantage of SIFT and rotation invariant uniform LTP performs favourably compared to SIFT and LBP, and computes more efficient than SIFT.