Human Bone fracture prognosis using Income inequality based Texture Feature and Support Vector Machine

Abstract

Bone fracture and related bone problems are most common throughout the world; people in every country are facing problems related to bone fracture. These are the prime reasons for bone fracture like due to some severe accident, or there may be chance that a person suffering from disease which weakens the bones like Osteoporosis or cancer. Therefore it is very much needed to quickly and accurately diagnose the affected area before giving any cure or treatment. Here we are proposing the technique through which we can detect and classify the fractured or healthy bones clearly, accurately and quickly. It works like a doctor’s tool and reduce his workload. Previous research work and data set is focused simply on classification of fractured bones but our research is capable of not only to classify and detect the fractured bone but the healthy bones as well, by considering data set consist of different types of human bones. The proposed approach analyzes the texture features for the bone diagnosis. In this regards, we performed performance analysis of the model using four GLCM (Gray Label Co-occurrence Matrix) texture features with and without Gini Index. The performance of the model using GLCM texture feature with Gini index is significantly improved. The proposed texture featured based SVM model have achieved the accuracy of 95% for the fracture bone.