A Complete System for Candidate Polyps Detection in Virtual Colonoscopy

TL;DR

This paper introduces a comprehensive computer-aided detection system for colonic polyps in virtual colonoscopy, achieving high sensitivity and low false positives, especially for larger polyps, through novel segmentation and feature analysis.

Contribution

It presents a complete pipeline combining segmentation, adaptive candidate delineation, and new texture and geometric features for improved polyp detection.

Findings

100% sensitivity for polyps >6mm with 0.9 false positives per case

93% sensitivity for polyps >3mm with 2.8 false positives per case

Effective detection of flat and small polyps in ground truth data

Abstract

Computer tomographic colonography, combined with computer-aided detection, is a promising emerging technique for colonic polyp analysis. We present a complete pipeline for polyp detection, starting with a simple colon segmentation technique that enhances polyps, followed by an adaptive-scale candidate polyp delineation and classification based on new texture and geometric features that consider both the information in the candidate polyp location and its immediate surrounding area. The proposed system is tested with ground truth data, including flat and small polyps which are hard to detect even with optical colonoscopy. For polyps larger than 6mm in size we achieve 100% sensitivity with just 0.9 false positives per case, and for polyps larger than 3mm in size we achieve 93% sensitivity with 2.8 false positives per case.