Colonoscopic Polyp Classification Using Local Shape and Texture Features

Pradipta Sasmal,M.K Bhuyan,Kunio Kasugai,Yuji Iwahori

doi:10.1109/access.2021.3092263

Abstract

In this paper, a method is proposed for colonic polyp classification which can perform a virtual biopsy for assessing the stage of malignancy in polyps. Geometry, texture, and colour of a polyp give sufficient cue of its nature. The proposed framework characterizes geometry or shape of a polyp by pyramid histogram of oriented gradient (PHOG) features. To encapsulate the texture of the polyp surface, a fractal weighted local binary pattern (FWLBP) descriptor is employed, which is robust to affine transformation. It is also partially robust to illumination variations which is generally encountered during endoscopy. The optimal feature fusion is done using a feature ranking algorithm based on fuzzy entropy. Finally, to evaluate the classification performance of the proposed model, kernel-based support vector machines (SVM) and RUSBoosted tree are used. Experimental results carried on two databases clearly indicate that the proposed method can be used in the colonoscopic polyps classification. The proposed method can give polyp classification accuracies of 90.12% and 84.1%, and AUC of 0.91 and 0.92 for publicly available database and our own database, respectively.

Highlights

C OLORECTAL cancer (CRC) is becoming a threat to the human race
Texture feature analysis is very important in polyp classification
The characteristics of the polyp may be perceived differently for different ambient conditions. To deal with this problem, spatial pyramid matching (SPM) and fractal dimension (FD) were incorporated with this framework

Summary

Introduction

C OLORECTAL cancer (CRC) is becoming a threat to the human race. It is the second leading cause of death globally [1]. Early diagnosis of cancer can increase survival rates Such cancer tissues are basically the polyps in their advanced stage, which are commonly seen in the lower abdominal region. The doctors resect the polyp from its stalk and do a comprehensive visual and histopathological analysis. Other work includes endoscopic video processing and 3-D reconstruction of polyps. These techniques involve both classical feature learning methods and deep feature learning techniques. Texture feature analysis is very important in polyp classification. Some of the papers related to texture feature analysis are studied in our work. Armi et al, [4] reviewed different texture descriptors for image analysis and classification. We are concerned about polyp classification, and literature study is confined only to this domain

Methods

Results

Conclusion