Development of a Virtual Reality Based Guiding System Incorporated with Facial Image Processing for Individuals with Spinal Cord Injury in Gait Training

It is reported from the previous studies in literature that individuals with Spinal Cord Injury (SCI) may not be able to regain their ability to walk, even though this is a primary goal of rehabilitation therapy. In fact, traditional lower extremity gait training may not be able to effectively, fully resolve complex gait activities for individuals with SCI. In recent years, an exoskeletal- or elliptical-assisted gait trainer has become an emerging novel technique for gait training on people with SCI and has shown its growing popularity in clinical applications. However, due to the poor cardiopulmonary function, people with SCI often encounter fatigue, or even fainted, during the training process. Since people with SCI may often experience fatigue and cardiopulmonary dysfunction during the assisted gait training, the patient may thus lose the motivation to perform the training. In this aspect, Virtual Reality (VR) can effectively incorporate an interactive learning process, and thus can provide a variety of attractive and challenging training in complex virtual environments. Therefore, in this project we propose a solution to the above problems. Our project aims to develop a facial expression detection algorithm and virtual reality guiding system, incorporating with the on-line monitoring of Heart Rate (HR), Heart Rate Variability (HRV), and Cardiac Stress Index (CSI), that can be applied to people with SCI during elliptical-assisted gait training in order to enhance the training motivation, to detect the exhaustion during the training on a timely basis, and to assess the cardiopulmonary function in a real-time manner for people with SCI. In this project, we expect to accomplish the following objectives in the upcoming three years: 1) to establish an interactive VR scenario/scene suitable for elliptical-assisted gait training, 2) to develop and realize an Artificial Intelligence-based (AI-based) facial expression recognition algorithm, and 3) with facial image processing based HR, HRV, and CSI estimates, to incorporate the bio-feedback mechanism, towards the integration of all the above three tasks/components with the other projects and then to finalize the system prototyping of the intelligent VR-based guiding and bio-monitoring suspension elliptical-assisted training platform.

Project ID:PB10901-0120
External Project ID:MOST107-2221-E182-011-MY3