Automated Bone Fracture Detection Using Deep Learning by a High-Accuracy Approach with CNN and DenseNet Architecture

Abstract

The people who are living in the world must stay healthy during their entire time span. Due to some reason, if their health conditions are damaged then they are going to face a lot of problems even managing their family. Additionally, if we see the health condition of people, then all their bones must be strong, so that they can move and run easily from one place to another without any hesitation. Once their bones are cracked due to some reason, and then may face a lot of problems. Here the authors have mainly focused on the identification of bone cracking which is a great issue in the world. Manually the bone cracking process is identified by the medical imaging technique X-Ray which is a very time- consuming process who are going to identified. As the people who are investigating regarding bone fracture, they may be busy with some other different tasks by their higher authority or by some other issues. Here we have implemented deep learning approaches like Convolutional Neural Network and an improved DenseNet architecture. All the validation process has been carried out by the most familiar dataset MURA. If we discuss about the outcome, then it has acquired as an approximation of 98.77 % which can improve the performance of machine learning approaches.

Ähnliche Arbeiten