A Version of Watershed Algorithm for Color Image Segmentation

Finding of semantic regions is the main objective of segmentation for image understanding. Automatic image segmentation is one of the major difficulties in the field of image processing. The watershed algorithm has some problems, like over segmentation, sensitive to noise, and high computational complexity. To overcome these problems, we have proposed a Modified Watershed (MWS) algorithm for color image segmentation by adaptively selecting threshold and masking mechanism over each color channel before combining the segmentation from each channel into the final one. For qualitative performance, proposed modified watershed algorithm performance compare with four modified watershed algorithms. For quantitative verification, proposed MWS method with two modified watershed algorithms in terms of executing times. The presented method have compared with FCM, RG and HKM techniques for color image Segmentation in 10 different classes of images with respect to PSNR, MSE, PSNRRGB, CQM and RFSIM. It is worth noticing that our proposed approach is low computational complexity. According to the visual and quantitative verifications, the proposed MWS algorithm is better than others three algorithms on the segmentation of color image.