Fully automatic Breast ultrasound image segmentation based on fuzzy cellular automata framework

Publication date: February 2018 Source:Biomedical Signal Processing and Control, Volume 40 Author(s): Yan Liu, Yong Chen, Bo Han, Yingtao Zhang, Xutang Zhang, Yanxin Su In this paper, an effective automatic image segmentation approach based on fuzzy cellular automata (FCA) framework is proposed for handling the uncertain ascription of pixels’ status during segmentation process. Different to the existing methods, the certain outputs of cell characteristics in traditional CA are transformed into the fuzzy decision expression. For selecting the seed pixels automatically, an automatic seed point templates generation approach is developed that employs a multiple thresholding fusion strategy by extracting different texture features based on intensity and non intensity distribution of pixels comprehensively. In its energy function, different image information comparison functions are modeled based on the complementarity of the local and the non local spatial information of image. The proposed method was performed on a series of clinic breast US images for studying its characteristic and validating its performance. Three overlapping area error metrics are used for measuring the performance statistically. Experimental results demonstrate that the proposed method can handle images with high speckle noise, complicate structures and blurry boundaries well, it is insensitive to initial condition, robust to noise and segments the ultrasound images accurately.
Source: Biomedical Signal Processing and Control - Category: Biomedical Science Source Type: research