Computer-Aided Detection and Diagnosis of Diaphyseal Femur Fracture

In the last decades, ameliorations in medical imaging technologies have enabled the advancement of computer-aided diagnosis systems to reduce the tasks of the doctors by early detection, screening out the easy cases, i.e., severity-based classification and in planning surgery, etc. In orthopedic surgery, computer-assisted diagnosis systems have obtained significances in automatically detecting and diagnosing fractures. In this paper, a system is presented which automatically detects and diagnoses the diaphyseal femur fracture part in the X-ray images using the combination of sliding window approach and support vector machine. Further, the back propagation neural network and probabilistic neural network classifiers are used to identify the type of diaphyseal femur fracture, namely transverse, spiral, and comminuted. The performance of this system is recorded with 175 real-patient data of abnormal and normal circumstances. The experimental results expose that the proposed method could be employed as an efficient tool to reveal the diaphyseal femur fracture automatically.