Last updated: September 03, 2008

CEB Projects

Print this   E-mail this

CervigramFinder: Cervigram Content Based Image Retrieval System

Project Member(s): Zhiyun (Jaylene) Xue, Sameer Antani, Rodney Long, George Thoma

The National Library of Medicine (NLM) and the National Cancer Institute (NCI) are creating a digital archive of 100,000 cervicographic images (cervigrams) and correlated diagnostic data obtained from the ALTS and Guanacaste studies . Access to this cervicographic database by image characteristics, such as color and texture of particular regions on the cervix, is expected to greatly benefit medical education and research into uterine cervical cancer, which is the second most common cancer affecting women worldwide.

CervigramFinder is a Content Based Image Retrieval (CBIR) tool with color, texture, shape, and location based CBIR functions that operate on a subset of the cervigram image collection. Given a query region marked by a user, the system can retrieve important cervix tissue regions that are similar to the query region with respect to image features listed above as well as any expert-annotated regions such as os, acetowhitened lesions (AW), squamo-columnar junction, columnar tissue, squamous metaplasia, mucus, and blood.

CervigramFinder is implemented as a Java Applet and is currently restricted to use by participating researchers or by invitation only. Please contact the developer to request access.

Registered users may click here to access the tool.

 Return to top