Location of Thesis Examination

Room 2009B Spencer Engineering Building

Degree

Master of Engineering Science

Program

Biomedical Engineering

Supervisor

Hanif Ladak

Abstract

Endoscopic cameras are increasingly employed in image-guidance procedures, where the video images must be registered to data from other modalities. However, such cameras are susceptible to distortions, requiring calibration before images can be used for registration, tracking and 3D reconstruction. Camera calibration is learned in a laboratory setting, where configuring and adjusting the physical setup is tedious and not necessarily conducive to learning. A centralized resource that utilizes 3D interactive components needs to be available for training on camera calibration. In this project, a web-based training environment for camera calibration is implemented called SimCAM. SimCAM was developed using the Web Graphics Library (WebGL), Open Computer Vision (OpenCV) library, and custom software components. WebGL and OpenCV were used to simulate camera distortions and the calibration task. The main contributions include the implementation and validation of SimCAM. SimCAM was validated with a content validity study, where it was found to be useful as an introduction to camera calibration. Future work involves improving the supporting material and implementing more features, such as uncertainty propogation.