A robust matching method for low-textured image based on co-occurrence probability of geometry-optimized pixel patterns

In this paper, we propose a template matching algorithm which is applicable for low-textured image like a range image. As for high-speed template matching, the Co-occurrence Probability-based Template Matching (CPTM) is a useful and effective method. This method uses some sets of selected pixel patterns that have relatively low occurrence probability in a template image. By using such a small number of distinctive data, reliable matching has been achieved in addition to high-speed processing. However, this method has a problem that extraction of distinctive pixels will be difficult when distribution of occurrence probability is uniform, for example, it is frequently appeared in range image. We improve the CPTM method for dealing with this problem. A key idea is to optimize geometric pixel relation in the pixel pattern when the proposed method calculates occurrence probability of pixel patterns. Experimental results have confirmed that the proposed method increase the detection rate from 73% to 90% without sacrificing its ability of high-speed. It means that performance of our method is prior to other conventional methods.