Anomaly Detection of Hallux Valgus using Plantar Pressure Data

Abstract

Machine learning is a superior tool that is unbiased and moderately comparable to the medical expert in making medical diagnostics if trained with correct supervision. In this paper we developed a supervised learning algorithm employing plantar pressure data to detect the anomaly called hallux valgus (HV) on a number of subject. Support vector machine (SVM) and its variants such as kernel SVM and ensemble SVM were evaluated on a plantar pressure open dataset. Results show that SVMs in general have the average classification rate of above 90 percent.