Automated classification of plantar pressure asymmetry during pathological gait using artificial neural network

Pathologies of the foot are amongst some of the most debilitating problems affecting individuals of all ages. Often, these pathologies are painful and correspond with high or abnormal plantar pressure, which can result in asymmetry between the feet during pathological gait. These problems, if left untreated, can escalate to severe plantar injury. The diagnosis of plantar pressure abnormalities can be unreliable, particularly when using the traditional methods based predominately on simple qualitative clinical screenings. It is therefore imperative for the early intervention and prevention of plantar injury, by reliably detecting plantar pressure abnormalities. This paper aims to evaluate the feasibility of applying an artificial neural network (ANN) in identifying, and correctly classifying plantar pressure asymmetry during control (healthy) and pathological gait. The results achieved using ANN classifier models applied to the plantar pressure asymmetry of 47 participants has demonstrated good network ability in differentiating healthy and pathological gait. The models' generalisation performance achieved classification accuracies between 87-100%. Such an automated foot pressure-based recognition model may prove to be useful for classification and diagnosis of other foot pathologies such as ulceration risk in the diabetic foot.