The Fractures Which Are Non-detectable And Which Are Not Diagnosed Properly Are The Two Major Challenges To The Orthopedicians And Radiologists In The Field Of Orthopaedics. In Many Countries The Count Of Patients With Bone-related Fracture Issues Is Increasing Very Rapidly, Especially Many Are Meeting With An Accident. Additionally, Due To A Shortage In Rural Areas Where The Medical Facility Is Very Poor And Unavailability Of Orthopaedics Makes The Introduction Of Computer-based Systems. In This Paper, A Computer-Aided Fracture Detection Technique Is Proposed Using X-rays Which Detect And Locate The Fractures Very Efficiently And Lessens The Burden Of Orthopedicians. Bag Of Visual Words (BOVW) Technique Is Proposed Where Features Are Labelled As Fractured And Non-fractured With The Training Of Back Propagation Neural Network (BPNN) Classifier With The Features Extracted From SIFT(Scale Invariant Feature Transform). The Proposed Method Experiments Over 300 Xrays Images Which Are Collected As A Dataset From KIMS Hospital, Hyderabad. The Experimental Results Give 90% Accuracy Compared To HTBFD (Hought Transform Based Fracture Detection And ANN ( Artificial Neural Network) Which Can Be Used For Better Location Of Fractures.
