Unsupervised color image segmentation is the one of the challenging topics. It simply segments an image into regions with respect to color and texture homogenity. Here, we propose an unsupervised colour image segmentation algorithm which uses both multiscale edge information and spatial colour content. The multiscale edge information is extracted using the multiresolution nature of dual- tree complex wavelet transform. Binary morphological operators are used to detect seed regions from the edge information. The segmentation of homogeneous regions is obtained using region growing followed by region merging in the CIE L¤a¤b¤ colour space. We also present an edge-preserving smoothing filter as a pre-process for the algorithm. We compare our algorithm with state-of-the-art algorithms and show the superiority of the proposed algorithm.
Bio:
Turgay Celik received the Ph.D. from Eastern Mediterranean University, Turkey. He joined the Temasek Laboratories at Nanyang Technological University, Singapore as Research Scientist in 2007. His research interests include multiresolution image processing, image sequence processing, and 3D computer vision.