An Efficient Camera Path Generation Method In Virtual Endoscopy

초록

Virtual endoscopy provides visualization of inner surface of the human organ using three dimensional volume models. Fast and accurate navigation is essential in virtual endoscopy. So we have to make a center-line by connecting points that have maximum distance from organ wall. Previous camera path generation methods have some problems to demand a lot of computation cost in preprocessing step and extra storage for spatial data structures such as a distance map or a potential field. We propose an efficient path generation algorithm using volume ray casting and the center of gravity in cross-section. We can get camera direction for the next frame identifying a ray that has maximum depth in current frame. Then camera position in the next frame is determined by using the center of gravity of organ’s cross-section. Our method constructs a camera path by doing this in every frame. In addition, it doesn't require preprocessing time and extra storage for maintaining spatial data structures.