Virtual stomach visualization and a stomach tumor detection system

Anatomy, the study of the structure of the human body is a fundamental of medical education. The main techniques and tools which are used in studying human anatomy are the traditional dissection of the human body and 2-dimensional diagrams in text books. However, a natural abhorrence towards dissection, emotional concerns, incapability of repeating the dissection, limitations in the number of cadavers and difficulty of capturing the real view of human anatomy presented in a text book plague traditional learning. This paper investigates the possibility of building an interactive Augmented Reality system which enables users such as medical students to practice dissecting a stomach with a great deal of freedom whilst enjoying a nearly real enhanced learning experience. The prototype which is developed using Goblin XNA, can be used as a learning tool which helps in traditional dissection and provides capability for identification of abnormalities in endoscopic images in which users can upload an endoscopic image and the system will automatically segment the abnormal area. Three image segmentation algorithms such as Watershed, Normalized Cuts and Topological derivatives are implemented using Matlab in order to find the best approach.