In medical images, the tube-formed tissues such as the blood vessels, the trachea, and the pancreatic duct are sometimes partially masked because of the constriction, stones in the vessels, the pancreatic cancer, etc. Therefore, it is not easy to automatically segment the region of tubes (ROTs) from medical images for visualizing the structures by using conventional image segmentation methods, because inference of ROTs is difficult. In this paper, we propose a fuzzy rule-based augmented reality method for finding non-continuous ROTs. We can obtain the ROT without extracting it. The physicians' procedure for finding the ROT can be eliminated by fuzzy inference techniques based on their knowledge. The employed knowledge is the intensity, the curve, and the radius of the ROTs. We apply the proposed method for finding the pancreatic duct from MR Cholangiography images. Through experimental results, we show that this method can successfully find the pancreatic duct from any data sets and it can clearly visualize the 3D shape of the ROT in MIP images.
展开▼
