Classification of Human Bones Using Deep Convolutional Neural Network

Abstract

In human body, there are total 206 types of different bones. Each bone has its own importance. It is very important to correctly identify human bone and then suggest treatment. To classify the human bones, we will use Musculoskeletal Radiographs (MURA) dataset. MURA dataset is one of the largest public radiographic image datasets. MURA dataset contains total 40,005 x-ray images of 14,052 patients, in which 36,808 images use as a training set and rest 3197 images use the testing set. These all images belong to seven different categories of bones such as finger, elbow, hand, forearm, humerus, wrist and shoulder. This paper aims to present a novel classification method using a deep convolutional neural network (DCNN). Dataset is freely available at https://stanfordmlgroup.github.io/competitions/mura.