A Survey on IMU-and-Vision-based Human Pose Estimation for Rehabilitation

With the rapid development of human pose estimation, the human-pose-estimation method applied in rehabilitation therapy has attracted more and more attention in today's aging society. Among them, inertial measurement units (IMU) and computer vision are two significant methods for human pose estimation. We mainly pay attention to the application of these two methods in the rehabilitation therapy field. In this paper, we first summarize the research status of IMU and computer vision in the field of rehabilitation therapy. Subsequently, we find that combining the two methods can produce better results. Finally, the main problems of the current research and possible development directions are identified.

1 Introduction

In recent years, the proportion of China's elderly population continues to grow. Central nervous system diseases such as multiple sclerosis, stroke and spinal cord injury, which are common in older people, often result in loss of upper and lower limb movement and sensory function. After these patients are discharged from hospital, they are often accompanied by regular physical health review and rehabilitation treatment [1], [2], which will lead to an increase in the demand for public medical service resources and medical costs for patients. Reducing “face-to-face” treatment is the best solution to the problem of healthcare resources and costs. Therefore, more and more people begin to pay attention to the development of family rehabilitation programs [3], [4]. The goal of rehabilitation therapy is to help patients regain physical strength and maintain good mobility [5], [6]. In fact, rehabilitation is a dynamic process that corrects any incorrect motor behavior through external devices in order to achieve the desired goal (such as ideal position) [7]. Therefore, human pose estimation becomes critical in home-based rehabilitation programs [8].