Content-Based Retrieval of Histopathological Images using Relational Tissue Graphs
|Sponsor: TUBITAK - Scientific and Technological Research Council of Turkey|
|Project no: 110E232|
|Principal investigator: Cigdem Gunduz Demir|
|Duration: Sep 2011 - Sep 2013|
In the current practice of medicine, histopathological examination is the routinely used method to diagnose several neoplastic diseases including cancer. Histopathological examination includes to examine a biopsy tissue under a microscope and to locate abnormal formations in the tissue. The reliability of this examination is closely related to the knowledge, expertise, and experience of a pathologist. The most important factor affecting his/her experience is the number of cases examined by the pathologist during his/her education and professional life and the diagnosis of these cases. Obviously, the experience of young pathologists or those that have not examined too many cases could be less and this may affect their decisions. Therefore, it is very important to provide the pathologists with cases similar to the case of interest together with their diagnosis.
The goal of this project is to design and implement an intelligent system that provides to search and retrieve cases visually similar to the case (image) of interest. This system aims at also retrieving cases that include a subimage causing the abnormality of the case of interest.