Faculty Advisor

Gennert, Michael A.


This report presents the research, methods, implementation, testing, and results of a system that detects patient respiratory motion in SPECT imaging using video streams captured using stereo optical web cameras. The main areas of focus were rectification of stereo image pairs, correlation of stereo features, and feature tracking. By comparing results from the feature tracking to data from patient respiration bellows, it was concluded that the system successfully tracked patient respiration. The system was implemented primarily in Java 2 version 1.4, using the Java Media Framework and JAMA Matrix Packages.


Worcester Polytechnic Institute

Date Accepted

January 2003


Computer Science

Project Type

Major Qualifying Project


Restricted-WPI community only

Advisor Department

Computer Science