Can machine learning decipher impairments from clinical gait graphs and reports?

Clinical gait analysis (CGA) involves the use of motion capture technology to assess the way a person walks. CGA is commonplace in paediatric and rehabilitation fields and provides highly accurate, objective information that is then used to optimise rehabilitation or intervention for people with disabilities. The strength in CGA is in the ability to provide objective data that directly informs treatment. However, the time that is taken to interpret data is problematic and so too is the subjective nature of observing data and applying a priority to which areas are of highest importance. With the explosion of artificial intelligence and the use of machine learning, there are many examples of large datasets being used to train models that are able to identify abnormalities. However, these are yet to be applied to CGA. The aim of this project is to use a large dataset of both quantitative biomechanical data and the accompanying clinical report to train a deep learning model to detect abnormalities/impairments in gait.