A system for colorectal tumor classification in magnifying endoscopic NBI images

Abstract

In this paper we propose a recognition system for classifying NBI images of colorectal tumors into three types (A, B, and C3) of structures of microvessels on the colorectal surface. These types have a strong correlation with histologic diagnosis: hyperplasias (HP), tubular adenomas (TA), and carcinomas with massive submucosal invasion (SM-m). Images are represented by Bag-of-features of the SIFT descriptors densely sampled on a grid, and then classified by an SVM with an RBF kernel. A dataset of 907 NBI images were used for experiments with 10-fold cross-validation, and recognition rate of 94.1% were obtained. © 2011 Springer-Verlag Berlin Heidelberg.