Nowadays a vast amount of people suffer from diabetes all over the world. The disease in its improved phase may cause ulcers especially in foots. To treat ulcers caused by diabetes, the region of ulcer and its area should be checked periodically by physician. Image processing algorithms can be used to segment the ulcer and estimate its area to assist the physician. In this thesis, two algorithms are developed for the purpose. In the frist algorithm, the operator roughly selects a few points of the ulcer boundary to set an initial contour, then localized region based active contour is used to segment the ulcers precisely. In the second algorithm, Hidden Markov Model and K Means Clustering are used to segment the ulcers roughly using the gray level image. The output of this step is used as the initial contour for a localized region based active contour to segment the ulcers precisely. The result of the first algorithm is better than the second one. The ulcer area is calculated after applying the first algorithm. Both algorithms are applied to a data base of 40 image and the ulcers were segmented and their area and perimeter were calculated. The results show that the average sensitivity, Dice similarity coefficient and positive predictive value for the first algorithm are 95%, 88/9% and 84/8% respectively and for the second algorithm are 79% and 68/6% and 69/33%. Keywords: Diabetic ulcers, Segmentation of wounds, K-Means clustering, Fuzzy-C-Means clustering, Susan edge detection , Active contours, Localized active contour, Hidden markov model.
