An efficient interactive segmentation algorithm using color correction for underwater images

Image segmentation is one of the basic strategies in image analytics. It is used for pattern detection, photo montage and object recognition applications. In underwater images light scrambling, suspended particles that exhibit in the water, marine snow, and color deviations etc. are some of the issues for image acquisition. Due to these issues, all the segmentation algorithms invariably produce poor results in the case of underwater images. We propose an efficient algorithm for extracting the foreground region from the underwater images based on a region of interest. Our segmentation procedure for underwater images includes color correction, contrast enhancement followed by an interactive GrabCut algorithm. Inclusion of color correction and contrast enhancement minimizes the number of touch ups required to enhance the segmentation when compared to GrabCut algorithm without these preprocessing steps. Our algorithm is efficient in the sense of time, when compared to the other algorithms for the same purpose.