This project developed an augmented reality (AR) finger rehabilitation system using vision technologies like OpenCV and MediaPipe, combined with gaming frameworks such as Pygame. The goal was to create a gaming interface controlled by hand and finger gestures, providing a form of rehabilitation for patients with limited hand mobility. The system tracks finger movements, collects motion data, and uses Pygame for its game interface. It was tested for accuracy and latency, with accuracy varying from 8.42% for the index finger to 11.11% for the thumb. The system's effectiveness was evaluated with a healthy control group, which showed non-significant changes after a 3-week training period. While the thumb showed the greatest motion variation (23.629°), no patients were tested, limiting conclusions about its rehabilitation benefits. The findings suggest that while the system is functional for data collection and gaming interaction, further improvements in accuracy and additional patient testing are needed for its use in rehabilitation.