Colon segmentation and the detection of colonic polyp with template matching in CT images

In this paper, we introduced a novel Computer Aided Detection (CAD) system for colonic polyp detection in CT data. The CAD system extracts colon region from CT images using cellular neural network (CNN) which its parameters of A,B and I templates are optimized by genetic algorithm in order to improve segmentation performance. Region of interest (ROI) of all slices were combined together to acquire a 3D ROI image and then we generate a 3D ROI image 3D segmented colon. Then the system performs 3D template matching with four layers of 12×12 cells to detect polyps. The CAD system was evaluated with 1148 CT images from 11 patients containing 15 marked polyps. The overall sensitivity of our CAD system is, 100% with the level of 10 FPs per case.