MT 18.01

Title: 

Automatic Detection and Segmentation of Shoulder Implants in X-ray Images

Author(s): 

Maya Belen Cervantes Gautschi Stark

Oral Defence Date: 

Wednesday, January 31, 2018 - 12:30

Location: 

TH 434

Committee: 

Assoc. Profs. Kaz Okada and Hui Yang and Prof. Dragutin Petkovic

Abstract: 

The procedures for repairing or replacing prostheses used in total shoulder arthroplasty (TSA) vary depending on the particular model of the prosthesis. If the model of the prosthesis is unknown, identification is performed by medical professionals based on visual inspection of X-ray images. This process is tedious and time consuming; indicating an unmet need for a tool that will aid surgeons in identifying prostheses quickly and accurately. A preliminary step towards the creation of such a classification tool is the segmentation of the prosthesis. This thesis describes the design and implementation of a software solution to the problem of detection and segmentation of TSA implants in X-ray images. The method implemented uses the Hough transform for circles to locate the implant, followed by segmentation using a seeded region growing method. Validation is performed by comparison with manually segmented ground-truth images, and by visual inspection of the results.

Keywords: 

Image processing, segmentation, object detection, X-ray images, total shoulder arthroplasty

Copyright: 

Maya Belen Cervantes Gautschi Stark