A MR-Mapping System for Guided Self-Correction in Postural Rehabilitation

For large-limb movement correction training scenarios, real-time correction and standard movement observation during the process are important requirements. Using computer vision to achieve real-time recognition of user postures and constructing personalized training feedback streams through dynamic interaction paths are novel channels to enhance the experience. In this study, we propose a Trainer-Anchored MR system for large-limb rehabilitation training, in order to achieve the goal of self-correction of movement learning with low cost and high engagement.