Research of image segmentation based on theory of morphological connectivity
Tianzhen Dong, Mingyang Qin, Wenju Li, Lanfeng Zhou, Qi Xiao · 2017 International Conference on Security, Pattern Analysis, and Cybernetics (SPAC) · 2017
Image segmentation is a common problem in image processing. Due to various image types and diversified applications, there is no widely accepted segmentation method. In this paper, the morphological connection theory is used to segment a color image into several irregularly shaped connected regions (i.e. equivalence class) based on their distributions of hue, intensity and texture, etc. As the pixels in a same connected region possesses the same or similar feature, the proposed method is provided with a good segmentation result for smooth and texture regions. The validity of proposed method is demonstrated through experiments.