Classical FE Analysis to Classify Parkinson’s Disease Patients

Parkinson’s disease (PD) is a neurodegenerative condition that affects the correct functioning of motor system in human body. Patients exhibit reduced capability to produce facial expressions (FEs) among different symptoms, namely hypomimia. Being so hard be detected its early stages, automatic systems can created help physicians assessing and screening patients using basic bio-markers. In this paper, we present several experiments where features are extracted from images FEs produced by PD healthy controls. Classical machine learning methods such as local binary patterns histograms oriented gradients used model images. Similarly, well-known classification method, support vector for discrimination between subjects. The most informative regions faces found with principal component analysis algorithm. Three were modeled: angry, happy, surprise. Good results obtained cases; however, happiness was one yielded better results, accuracies up 80.4%. paper classical research community; their main advantage they provide clear interpretability, which valuable many researchers especially clinicians. This work considered good baseline motivates other propose new methodologies yield while keep characteristic providing interpretability.