Anterior Cruciate Ligament (ACL) Coronal View Injury Diagnosis System using Convolutional Neural Network

Abstract

ACL injury is one of the most common injuries in sports activities or events. Failure to detect it would endanger the athletes' future. In this research, knee joint magnetic resonance imaging (MRI) is studied for the development of a computer-aided system to classify ACL injury. This work aims to develop a deep learning system applying Convolutional Neural Network (CNN) with Confusion Matrix analysis to assist medical experts in making decisions regarding the types of an ACL knee injury in the form of a classification based on complete tear (CT), a partial tear (PT) and normal or non-injury classes. 360 knee MRI images (coronal view) were used to develop an alternative feature extraction and classification technique in deep learning as compared to existing automated system. The result of confusion matrix analysis accuracy of the classification of ACL injury is 94.7%.