Share Email Print
cover

Proceedings Paper

Segmentation techniques for intestinal lumen detection for endoscopic color images
Author(s): Marta Patricia Tjoa; S. M. Krishnan; Shunren Xia
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Use of electronic video endoscope for viewing the lumen by clinicians is frequently carried out. Computer-assisted for processing and analysis the images will aid non-expert clinicians in diagnosis. Abnormalities can be detected by characterizing the features of the segmented endoscopic images. A new technique has been developed for segmenting the endoscopic image obtained from the large intestine using color parameters for detecting the lumen. This novel method for endoscopic image segmentation exploits the local homogeneity index (?) definition of the pixels in a given neighborhood of the image. The proposed segmentation is realized in two stages. Initially, segmentation is performed using ?, which is measured in terms of standard deviation and discontinuity in the achromatic distribution of the region. A modified peak-finding algorithm is employed to segment the image from the corresponding histogram. In the second stage, the regions obtained in the first stage are divided into subregions based on ? in the chromatic domain. Merging the region using color difference measures alleviates the problem of oversegmentation. This novel method is compared with a minimum variance region-growing method. The proposed technique is tested using clinically obtained colonoscopic images. The preliminary results indicated that both techniques are feasible. The technique based on ? showed better performance than a minimum variance region-growing method because it includes the analysis of both local and global information by using ? histogram. The proposed scheme can be adopted for computer-based analysis of the endoscopic information to facilitate early detection of colorectal cancer.

Paper Details

Date Published: 13 September 2002
PDF: 8 pages
Proc. SPIE 4922, Color Science and Imaging Technologies, (13 September 2002); doi: 10.1117/12.483118
Show Author Affiliations
Marta Patricia Tjoa, Nanyang Technological Univ. (Singapore)
S. M. Krishnan, Nanyang Technological Univ. (Singapore)
Shunren Xia, Zhejiang Univ. (China)


Published in SPIE Proceedings Vol. 4922:
Color Science and Imaging Technologies
Dazun Zhao; Ming Ronnier Luo; Kiyoharu Aizawa, Editor(s)

© SPIE. Terms of Use
Back to Top