Selection of suitable segmentation technique based on image quality metrics

Image segmentation [Gonzalez RG, Woods RE. Third edition digital image processing. South Asia: Pearson; 2014, Sridhar S. Digital image processing. India: Oxford; 2011] is a very important step in applications of the image processing field. Segmentation will divide the image into different parts and extract a required area known as Region of Interest (ROI). ROI varies with applications. For the effective implementation of image processing applications segmentation of the required area is the first major problem. There is no single universal segmentation algorithm to extract ROI from all types of images. So the best solution is to try all algorithms and pick a suitable one for a particular image. In this paper, we compare different image segmentation algorithms by applying them to different images and by calculating various image quality parameters such as MSE (Mean Squared Error), SNR (Signal to Noise Ratio), PSNR (Peak-Signal to Noise Ratio), EPI (Edge Preservation Index) and SSIM (Structural Similarity Index Metric).