Foot Arch Classification via ML-based Image Classification

Foot pain have become one of the common health problem. One of the commonly-used noninvasive method to relieve foot pain is to insert insoles in ones’ shoes. However, choosing the right insoles strongly depends on foot arch types, i.e., high arch, normal arch, and flat foot. Aside from manual classification, using foot images become an alternative methods to classify the foot type. We propose to develop mathematical models using machine learning techniques to improve the accuracy and reduce the time of the foot arch classification from foot pressure scanning image. 200 foot images were used to develop the models by applying decision tree, random forest, support vector machine, artificial neural network, and XGBoost algorithm. We found that the decision tree classifier with the features including different areas of part of foots, arch index, whole foot area, and side of foot has the best performance than the other classifiers in terms of accuracy, precision, recall, F1 score, and the number of features. The results also demonstrates that the obtained model can classify foot arch types with high accuracy at 95% on the testing experiment.