Hand and Fingertip Detection for Game-Based Hand Rehabilitation

Hand rehabilitation is the process of recovering hand movements to return to normal. For rehabilitation to be effective, it is necessary for patients to practice repetitive movements. During the process, the patients usually feel bored, lack motivation and enjoyable time. There is a growing interest in building games for rehabilitation to make it more interesting and motivating for patients. However, the major obstacle is that the devices are expensive, and the patients must go to use them in hospitals. This paper focuses on designing and implementing hand rehabilitation software by using hand and fingertip detection. The purpose of the project is to make the rehabilitation process more accessible, enjoyable, affordable, and can be made portable for patients to use at home. Deep learning techniques were utilized to perform hand and fingertip detection. The SuperFox game is developed; the game is controlled by commands associated with hand and fingertip movements. The system was tested with 10 participants to evaluate effectiveness of the game controls. The results showed that the system is feasible in controlling the game and able to be used to create motivation and enjoyable time for patients.