Detection of Acetabulum Fractures Using X-Ray Imaging and Processing Methods Focused on Noisy Images

Acetabulum fractures are a clinical problem that primarily affects people between the ages of 21 and 30. The incidence of this type of injury has increased due to accidents involving both immediate and late implications. The analysis of an x-ray image is the first step in evaluating this type of injury and determining the type of treatment the patient should undergo. Identifying if the traumatic injury is located in the pelvis, is an acetabular fracture or not, serves as a support for all those orthopedic surgeons who have to deal with this type of injury. A method of X-ray image processing was proposed using the Local Binary Pattern (LBP) feature extractor and the Support Vector Machine (SVM) classification method. Also, the use of pre-processing methods to improve image quality is a key contribution. The results show 80% accuracy in classification using the CLAHE preprocessing method.