LAUSR

In today’s era of new advancements, diagnosing a pathology at an early stage has given rise to the development of automated diagnostic systems. Knee Osteoarthritis (KOA) being among one of the most painful joint disorders is the root cause for disability, particularly in elderly population. Gait based recognition of KOA is a prominent area that requires deliberations from the end of researchers, academicians and scientists to develop more automated systems that not only offer reliability and accuracy but are also affordable for common man. This article aims to provide an in-depth investigation of efforts directed towards vision-based, sensor-based and hybrid KOA identification. The study is based on the historical data gathered and background obtained viz-a-viz clinical gait analysis. An extensive survey of KOA gait acquisition modalities and feature representation approaches for the purpose of critically examining them are also presented. The study surveys the statistical metrics used for evaluating KOA, considering relevant articles. Based on the survey, this article aims to provide an up-to-date review of machine learning techniques for classification of KOA and healthy subjects. Furthermore, this article also identifies open research challenges existing in the literature that could be explored further for providing more effective KOA analysis. Finally, this article presents the future perspectives and provides an outline of the proposed work for efficient KOA diagnosis based on vision-based gait.