Computer-Aided Diagnosis System for Bone Fracture Detection and Classification: A Review on Deep Learning Techniques

Bone fracture detection and classification was a large discussed topic over the last few years and many researchers proposed different technological solutions to tackle this task. Despite this, a universal approach able to support the classification of fractures in the human body still does not exist today. We aim to provide a first discussion concerning a selection of research works done in the technological domain, with a specific focus on Deep Learning. The objective was to underline a picture on the most promising studies for stimulating a knowledge improvement in the specific focus of bone fracture classification, necessary to start the development of an optimal shared framework. The evaluation has been made involving a first qualitative assessment based on strengths and weaknesses, providing a usage scenario evaluation. This could support the development of a helpful Computer Aided Diagnosis (CAD) system able to drive doctors in diagnosis tasks reducing diagnosis time, especially in the most complex tasks, and supporting the reduction of wrong diagnosis issues, especially during stressful working conditions, as what frequently happens in many emergency departments.