Seeded region growing segmentation on ultrasound image using particle swarm optimization

Abstract

Ultrasound imaging is one of the most popular and cheapest noninvasive medical scans. At the time of image acquisition, there may be degradation in the quality of image in the form of speckle noise. In recent times, many researches have made various experiments to enhance the quality of medical imaging. However, there is scope to further enhance it. In the proposed method, finding out the seed pixel randomly is the basic problem, which is treated as an optimization problem. This problem can be solved by Particle Swarm Optimization. Using Particle Swarm Optimization algorithm, the fitness function can give us the appropriate seed pixel for the desired ultrasound imaging. In this paper, a novel method is proposed, wherein segmentation will be applied on a fuzzy filtered image. The fuzzy filter applies fuzzy rules to detect regions in the image viz. edge region, homogeneous region, and noisy region by using different gradients, and then filters the noisy region using fuzzy membership rules. The proposed method has been tested on different ultrasound images, and the experimental results demonstrate its effectiveness.