Today medical instruments assists specialists to diagnosis and treatment of inside injuries. One of the most common methods for bone injuries diagnosis is X-ray Imaging. Low quality of some of this images and carcinogenesis of X-ray challenge diagnosis procedure. More over psycho visual limitation leads to lack of diagnosis of some bone fractures. Therefore computer aided fracture detection can improve diagnosis procedure. Challenging structure and quality variety of bone X-ray images cause to lack of existence of effective segmentation and fracture detection methods. In this thesis we will propose some methods for quality enhancement, bone segmentation and fracture detection in long bone X-ray images. All of this methods are based on an stable feature of long bone which is robust against destructive effects. Evaluation of results show that the accuracy of proposed methods are better than previous methods.
