Modified deep learning ResNeXt model for Human long bone fracture detection and classification

Abstract

The field of medical science is going to take advantage of Machine learning. It has increased dramatically over the last decade. Nowadays, you can see other innovations used in medical sciences, such as machine learning and deep learning. They can help to diagnose the illness or cause. It can also aid in the healing process by keeping notes. At a similar pace, an upper hand has been provided to the physicians for image processing by incorporating computers. Bone fractures are normal these days, and the identification of fractures is a critical part of orthopedic X-ray imaging. The automated technique lets the doctor quickly begin medical treatment. Using Machine Learning and CNN (Convolutional Neural Network), we suggest a new deep learning model perform bone diagnosis by eliminating discontinuity followed by segmentation of the image in a system that detects bone fractures. It overcomes the shortcomings of the previous approach that operates only on examination of the texture features. The proposed deep learning modified ResNeXt model performs much better than the state-of arts.