Polyps's region of interest detection in colonoscopy images by using clustering segmentation and region growing

In this paper we propose a novel method to automatically detect the polyp's ROI (Region Of Interest) based on a Gaussian Mixture Model, Esperance Maximization algorithm segmentation, Hough transform and region growing. Our main objective through this work is to help in the early detection and precise analyze for the polyps diagnosis, which is a huge apport for the health care. The experimental results show that the proposed method can achieve 83.15% accuracy and 98% sensitivity on our data set.