Patellar fracture analysis using segmentation and global thresholding techniques

Abstract

Radiologists identify abnormal pathologies including fractures with a high level of accuracy. However in some cases the examining reader accuracy may show high miss rate while reading X-rays containing abnormalities such as multiple patellar fracture. Accurate diagnosis of fractures is vital to the effective management of patient injuries. As a result, detection of patellar fracture is an important orthopedics and radiologic problem. In this paper, attempt has been made to develop an algorithm which will identify global thresholding ranges for different edge detection operators e.g. Sobel, Prewitt, Canny, Laplacian of Gaussian for analysis of fractured patella, which will help the orthopedic surgeons for analyzing the fracture in a better form than conventional method of diagnosis which is subjective, time consuming and tedious. The processing algorithms are developed on MATLAB 7.6.0 (R2008a) programming platform. © 2011 Springer-Verlag.