L'evy-flight firefly algorithm based active contour model for medical image segmentation

Image segmentation is the pre-processing task for many computer assisted medical imaging applications. Recently, the Active Contour Models (ACMs) have shown their superiority over traditional low level techniques in segmenting ill-defined medical images. Irrespective of continual development and refinement of these models for the last decades, sensitivity to the initial position of contour and the entrapment within local minima are among the problems inflicted on ACMs. ACM based segmentation is an optimization problem in the sense the objective is to search for a contour with minimal energy. The L'evy flight Firefly Algorithm (LFA) is a nature-inspired meta-heuristic algorithm that is very powerful in solving global optimization problems. In this paper a hybrid approach that integrates ACM with LFA in order to improve its segmentation ability. To validate the performance of the proposed approach, experiments have been conducted on the real CT images of abdomen. The results show the ascendancy of the proposed approach in reaching contour concavities as compared to the traditional ACM.