Analysis of Pose-Driven Computer Vision Algorithm for Tremor Differentiation

This pilot study aims to evaluate the efficacy of a computer vision-based system developed to differentiate between rest tremor, essential tremor, and cerebellar tremor. The study will involve a total of 75 subjects, with 25 participants in each tremor category, selected from various age groups above 18 years. Each subject will be independently evaluated using the developed computer vision system. The system utilizes video recordings of the patients’ hand and limb movements, focusing on capturing tremor activity during standardized postures. These include the rest pose (with hands placed on the knees while seated), the stretched hand pose (arms extended parallel to the ground), and intentional movement tasks (such as finger-to-nose or object-reaching movements). Each participant is required to sit for a maximum of 10 minutes to complete the evaluation individually. The recorded videos are processed using the developed computer vision algorithms to extract key motion features such as tremor frequency and trajectory patterns. These features are then analyzed to differentiate between Parkinsonian tremor, essential tremor, and cerebellar tremor.