Hierarchical classifiers for detection of fractures in x-ray images

Fracture of the bone is a very serious medical condition. In clinical practice, a tired radiologist has been found to miss fracture cases after looking through many images containing healthy bones. Computer detection of fractures can assist the doctors by flagging suspicious cases for closer examinations and thus improve the timeliness and accuracy of their diagnosis. This paper presents a new divide-and-conquer approach for fracture detection by partitioning the problem into smaller sub-problems in SVM's kernel space, and training an SVM to specialize in solving each sub-problem. As the sub-problems are easier to solve than the whole problem, a hierarchy of SVMs performs better than an individual SVM that solves the whole problem. Compared to existing methods, this approach enhances the accuracy and reliability of SVMs.