Bone Fracture Detection

The bone is a major component of the human body.   Bone provides the ability to move the body. The bone fractures are common in the human body. The doctors use the X-ray image to diagnose the fractured bone. The manual fracture detection technique is very time consuming and also error probability chance is high. Therefore, an automated system needs to develop to diagnose the fractured bone.  For our project we have used MURA dataset which was released by Stanford university in year 2017.

The x-ray is used in the computer-based system to perform the fracture bone diagnosis. The bone image contains noise. Therefore, a suitable preprocessing algorithm is used to remove noise and edges in the image. After that feature are extracted from the bone image. Finally, system is trained with the features and classification is performed by the ML (machine learning) algorithms.

Bone fracture detection and segmentation will be done by  X – ray images. First, data containing images of X-Ray have to be prepared. Then this image will be divided into two parts. 1. For training. 2. For validation. 80% of images should be kept in the training folder. After that Manipulation of the dataset will be done and dataset will be converted in well-defined manner and then Convolutional neural network is to be applied after creating training model respectively. After this, train images will be molded. Then the model should be trained. At last, testing it by adding random image of X-Ray will complete the task.  After that confusion matrix of trained model is created and then kappa score is predicted in this project kappa score is 0.6800 and ROC curve is also generated for batter understanding of the system.

This Project can be improved to the level where we can get more accurate results. And also, by launching this project as web application we can improve its usefulness and make it as easy to use.