Enhanced and Effective Computerized Classification of X-Ray Images

According to the World health organization (WHO) accidents became very common. Due to the increase in accidents automating of fracture detections became very essential. To detect the fracture it is necessary to classify the X-ray image first. This paper proposed an effective methodology for the classification of x-ray images to classify the automated explanation to provide efficient and effective results to the physicians and radiologists for making a decision. An attempt was made and a framework presented in this paper, which involves images being pre-processed using M3 filter for denoising, segmentation by K-means clustering, preceded by Statistical feature extraction. Classification of X-ray images are categorized into chest, spine, foot, palm, skull, the head is carried out by comparing the K-Nearest Neighbour (KNN) algorithm, Support Vector Machine(SVM) and Back Propagation Neural Network(BPNN). Overall 88% of accuracy in the analysis of X-ray images is acquired in BPNN Compared to other techniques.