Algorithm for detection of head tremor according to data of a smartphone video camera of a biomedical monitoring system

Modern conditions demand active digitization from humanity across various spheres of activity and daily life, facilitating faster task completion and simplifying processes. Self-diagnosis allows individuals to identify symptoms, which can serve as a basis for consulting medical professionals, especially crucial in critical situations where lives are at stake. Thus, it is clear that the development of such systems is a relevant challenge. In this context, head tremor plays a significant role as it may indicate the presence of Parkinson’s disease or multiple sclerosis. The aim of this work is to develop a head tremor detection module suitable for integration into smartphone applications. The study employs a method based on analyzing data from the optical sensor, namely the front camera of the smartphone. This method utilizes an open machine learning model, ML Kit, for facial recognition, along with a specially designed algorithm for processing results. Testing demonstrated an accuracy of 0.92 according to the accuracy metric. This approach offers a novel method for detecting head tremors and highlights the effectiveness of using ML Kit’s standard model for similar tasks on smartphones, which can also be applied within a larger biomedical diagnostic system.