Study of Measuring Motor Function in Dementia by Tapping Task Using IoT and Image Recognition

Dementia has emerged as a significant issue in contemporary Japanese society. Assessing finger motor function is pivotal for diagnosing and rehabilitating individuals with dementia. In this study, we design a web application to gauge manual dexterity through tapping exercises. Our fundamental concept combines finger identification using a web camera and tactile panel input to quantify finger actions. In our proposed methodology, we amalgamate real-time visuals captured by a web camera into the hand.js module within the MediaPipe framework, autonomously generating 21 critical finger data points. Concurrently, we establish touch panel control buttons on the web interface. In the context of the operating interface, participants manipulate their fingers in sync with numbers or features displayed every two seconds and utilize touch panel input to log button presses. This approach facilitates an all-encompassing assessment of finger motor aptitude.