Automated Nonlinear Feature Generation and Classification of Foot Pressure Lesions

Plantar lesions induced by biomechanical dysfunction pose a considerable socioeconomic health care challenge, and failure to detect lesions early can have significant effects on patient prognoses. Most of the previous works on plantar lesion identification employed the analysis of biomechanical microenvironment variables like pressure and thermal fields. This paper focuses on foot kinematics and applies kernel principal component analysis (KPCA) for nonlinear dimensionality reduction of features, followed by Fisher's linear discriminant analysis for the classification of patients with different types of foot lesions, in order to establish an association between foot motion and lesion formation. Performance comparisons are made using leave-one-out cross-validation. Results show that the proposed method can lead to $\sim 94$% correct classification rates, with a reduction of feature dimensionality from 2100 to 46, without any manual preprocessing or elaborate feature extraction methods. The results imply that foot kinematics contain information that is highly relevant to pathology classification and also that the nonlinear KPCA approach has considerable power in unraveling abstract biomechanical features into a relatively low-dimensional pathology-relevant space.

I. Introduction

Application of pattern analysis and machine learning to biomechanics and human gait is important for the automated diagnosis of many pathologies related to kinesiological debilitation and the evaluation of treatment regimes. Such algorithms can automatically process measurements obtained from most modern sensors, which are otherwise very cumbersome to analyze with traditional techniques due to factors related to high dimensionality and small samples of the datasets, temporal dependencies, as well as intersubject and intertrial variabilities, data redundancies, and nonstationarities and nonlinearities of the signals [1], [2]. Some recent representative examples of advanced techniques in biomechanics include the use of kernel methods for the classification of age from gait [3], classification of foot lesions using feature selection from discrete kinematic gait events [4], statistical approaches for foot pressure imagery [5], frequency analysis for automatic activity detection [6], and adapting neural networks for the estimation of gait kinematics from wearable sensors [7].