In Partial Fulfillment of the Requirements for the Degree of Master of Science

Adrien Sitter

will defend his thesis

Registration and Visualization of 3D Breast Images

Abstract

Breast cancer is one of the most common invasive cancer in women. Current treatment for women diagnosed with this cancer often consists of surgical removal of the entire breast (i.e. mastectomy). Most women who choose to undergo mastectomy, however, are concerned about their body image and reconstructive surgery is a potential option for them to restore the body appearance. It is well understood that reconstructive surgery has an important psychological impact on the patients as it contributes to the patients' self-perception of their body. A critical step in this process is for the patients to be able to view their recons tructed breast prior to election of the procedure. In Houston, Texas, psychologists, plastic surgeons, engineers and professors have initiated the Breast Aesthetics project and have gathered together their efforts for the development of a 3D computer technology to help patients actively participate in the rebuilding of their breast and facilitating the choice of reconstruction options that would best meet their quality-of-life concerns. This is done by offering patients a 3D view of the changes of their breast during the reconstruction process and simulation of what their breast would look like after reconstructive surgery.

This thesis describes the Breast Reconstruction Visualization software. It is presently being used by plastic surgeons and is able to provide 3D Visualization of a patient's breast 3D pictures. It also provides several measurements by using adequately selected fiducial markers on those pictures. The overall software architecture and de sign of the software is presented. In addition, visualizing changes to the breast over multiple stages of reconstruction is an important feature for the patient, and this require registration of 3D pictures. This thesis details the algorithms and implementation of fully automated 3D registration in three steps. The first step is the detection of two landmarks on the body of the patient: the Sternal Notch and the Umbilicus. Next, thirteen points are automatically selected on the body of the patient based on the positions of the landmarks. The final step is a rigid registration based on the previously detected and selected points.

Date: Tuesday, July 28, 2015
Time: 10:00 AM
Place: PGH 550
Advisor: Dr. Shishir Shah

Faculty, students, and the general public are invited.