Identification of Parkinson's Disease Through Facial Image Classification: A Systematic Review

Abstract

Parkinson's Disease (PD) is one of the most prevalent neurological conditions, affecting a significant portion of the elderly population worldwide. Characterized by motor symptoms such as tremors, rigidity, and bradykinesia, as well as non-motor symptoms like hypomimia (reduced facial expressiveness), PD presents significant diagnostic challenges, especially in its early stages. In this context, this systematic review (SR) aims to explore machine learning and deep learning techniques in detecting PD through facial image analysis. The study identifies the key works in the field, the computational techniques employed, and the datasets used. The main findings suggest that facial phenotyping can be a strong indicator for the early diagnosis of PD. Various approaches, including Convolutional Neural Networks (CNNs) and other deep learning models, have shown promise in detecting subtle changes in facial expressions associated with hypomimia. Integrating multimodal data sources, such as voice recordings, hand-drawn sketches, movement analysis, and facial images, has also demonstrated the potential to enhance diagnostic accuracy. This review highlights the potential of facial image analysis as an innovative approach for the early detection of PD, which could lead to more timely and effective interventions.