Design and implementation of automatic hand's pose classification systems for multiple sclerosis patients perfoming the Nine-Hole Peg Test

Neurological damage in Multiple Sclerosis (MS) affects the motor, sensory, visual, cognitive and sphincteric systems, resulting in different degrees of disability. A coredeficit affecting multiple sclerosis patients is the upper limb dysfunction. Since this impairment is not properly recorded in the disability scale of MS, Expanded Disability Status Scale (EDSS), other tests have been developed. In this direction, the Nine Hole Peg Test (9HPT) is widely considered a gold standard metric for manual dexterity. However, its quantification strongly depends on the subjectivity of each clinician. This Master Thesis describes a developed framework to automate and objectivize the 9HPT. The framework compares the performance of different Machine Learning and Deep Learning models on MS diagnosing. To that end, two different experiments are conducted. Firstly, a system is built to analyze videos of subjects performing the 9HPT, to extract the hand’s pose using Convolutional Neural Networks, to acquire the Hand’s kinematic features and to finally classify the severity of the disease based on several Machine Learning algorithms. For the framework validation, two classification approaches are proposed. The first one is a binary classification problem, separating the subjects between control and patients. The second one is a multiclass classification problem, where the patients are separated into three different classes based on the severity assessed by the clinicians. A technical validation has been implemented with an accuracy of up to 96.67% for the binary classification and 92.22% for the multiclass. Secondly, the multivariate time series data representing the hand movement when performing the 9HPT is classified using several Deep Learning architectures. In this direction, a binary classification problem is solved to detect if the subject suffers of MS and an accuracy of 88.24% is obtained