Rough Set Theory-Based Image Segmentation: A Comparison of Approaches in Two Color Spaces
Alberto J. Patlan-Rosales, Raúl E. Sánchez-Yáñez · 2012
This work presents an evaluation of color image segmentation based on rough set theory. A performance comparison of two algorithms in different color spaces, RGB and CIELUV, is carried on. In this histogram-based approach to segmentation, the concept of Histon plays a fundamental role. Thresholds are obtained using a roughness measure, and the segmentation is accomplished using a region merging procedure. Test series using a standard database are performed. Here, a quantitative measure of similarity between an original image and the segmented one is used for evaluating the outcomes. According to these results, we conclude that segmenting for the methodology in RGB is more recommendable than segmenting using the methodology for the CIELUV color space, at least for the rough set-based implementations considered for this study.