Enhanced Color Image Segmentation of Foreground Region using Particle Swarm Optimization

This paper proposes a new segmentation approach which aims to segment only the foreground of an image after background elimination. Background elimination is treated as an optimization problem and is solved by using principle of PSO. The proposed algorithm is a thresholding method used to eliminate background from an image assuming that the image to be threshold contains two classes of pixels or bi-modal histogram(foreground and background). This gives a low level binary representation to the image eliminating the background and highlighting the foreground part. Based on the distance and similarity among the connected components in the binary image, it is segmented and a different similar color is assigned to each of the segment to preserve the color information contained in the real color image