This paper evaluates the performance of an automated fly-through method of virtual colonoscopy based on the rate of undisplayed regions. When performing a diagnosis using a virtual colonoscopy (VC) system, the doctor observes the state of the colonic wall. Automated fly-through is often used to reduce the operation load on the virtual endoscopy system (VES). The system generates automated fly-through paths based on medial axes extracted from colon regions. There are many folds (i. e., “haustra folds”) in the colon, and these folds create blind regions in the images obtained during VC. There have been no studies on the rate of undisplayed regions resulting from occlusion during VC. This paper quantitatively evaluates automated fly-through paths generated during VC based on a method that measures the rate of undisplayed regions over the whole colon area in VES. We generated various fly-through paths and measured the rate of undisplayed regions. Experiments using three abdominal CT images showed that approximately 30% of the colon area was not displayed at all during automated fly-through.
