The developed system is the first AI-based rehabilitation system in the world that dedicated to automatically monitor and quantitatively analyze the rehabilitation progress under the LSVT BIG and re-abling for Parkinson＇s disease. Our system facilitated rehabilitation in: (1) patient compliance at home, (2) the minimum length for each rehabilitation to take effect, and (3) tracking and analyzing the rehabilitation process.
The core technologies included: (1) using webcam to capture patient’s skeleton, measure posture and to calculate angles of various movements; (2) developing wearable sensors to assist some motions that are invisible to the webcam; (3) designing an AI model to evaluate the rehabilitation scores; (4) integrating the Unity 3D gaming platform to instantly show the patient＇s gestures, allowing all personnel involved to gauge the effectiveness of the proposed treatments. The developed system is effective to quantitatively evaluate the rehabilitation progress.

The developed system broke the clinical limitation by proposing a novel method to monitor every exercise, angle, conformity to prescribed program for PD rehabilitation. The system allows practitioners to monitor the rehabilitation progress and make any necessary adjustments accordingly. Integrating with AI and big data analytics, this system achieves a giant step towards precision healthcare. Due to the COVID-19 pandemic many elderly cannot go to hospitals for rehabilitation. The proposed system provided an effective option for rehabilitation at home. The presented work also developed an interactive interface for the physical therapists or physicians to remotely monitor each rehabilitation session as done at the hospitals to assure the effectiveness of re-abling rehabilitation.

The proposed system is not limited to monitor rehabilitation progress on PD or re-abling patients. It can be applied to patient＇s rehabilitation anywhere and anytime. This system is especially suitable for patients due to COVID-19 pandemic that block down to the hospitals. A patient can do exercises or practice any rehabilitation session at home. A patient can watch his own and other three patients motions on their own segmented screens. Practitioners can remotely monitor the rehabilitation procedures, track the progress and make any necessary adjustments accordingly. The system can be extensively used in the realm of multi-reabling healthcare 2.0, community healthcare, etc. Experimental results verified that the proposed system is effective and comparable with the professional equipment.