Computer-aided teniae coli detection using height maps from computed tomographic colonography Images

Monday, October 24, 2011 — Poster Session I
Noon – 2:00 p.m.	Natcher Conference Center	CC	BIOINFO-18

* FARE Award Winner

Authors

• Z Wei
• J Yao
• S Wang
• R Summers

Abstract

Computed tomographic colonography (CTC) is a minimally invasive technique for colonic polyps and cancer screening. Teniae coli are important landmarks on the colon surface and serve as references for guiding virtual colon navigation and colonic polyp registration. We propose a novel method for teniae detection on CTC, which consists of four steps. The first step is to obtain a 2-D height map from a 3-D colon. The 3-D colon surface is unfolded and stretched to a rectangular region, then projected onto a 2-D height map, which records the elevation of the colon surface. In the second step, the height maps are sent into 2-D Gabor filter banks, where images are characterized by their responses to a set of orientations and special-frequency selective Gabor filters. The third step identifies fold centers with local maxima and thresholding. The final step is to extract teniae, which are located where the haustral folds meet. We tested our algorithm on a CTC dataset with seven cases. Root mean square error (RMSE) was employed for assessment. We normalized RMSE with the circumference of colon to obtain a percentage error rate. The proposed method achieved an average normalized RMSE of 5.66%, with a standard deviation of 4.79%.

back to top