54 2026 TSMA physiotherapists, and strength trainers to grasp individual data more accurately, thus creating personalized training and feedback. This article will expand on three major themes: markerless motion capture, wearable sensors, and AI-assisted analysis and prediction. It will explain how these technologies are being practically applied in competitive sports, fitness training, injury prevention, and health promotion—collectively reshaping the future landscape of the sports industry. 1. Markerless Motion Capture: The Key Turning Point Beyond Tradition Traditional motion analysis relied on expensive, bulky, and complex ‘marker-based systems,’ such as Vicon or Qualisys. These required reflective markers to be attached to multiple parts of the subject’s body, while multiple infrared cameras captured three-dimensional dynamic data. Although highly precise, they were restricted by usage space, lighting, and operational expertise, making them suitable only for laboratory environments and difficult to extend to general applications. The birth of ‘markerless motion capture’ technology aimed to overcome these limitations. This technology integrates computer vision with deep learning algorithms, and only requires a simple RGB camera (even a smartphone camera) to capture motion trajectories, joint angles, and spatial positions, reconstructing an accurate 3D motion model. In the past, motion analysis could only be conducted in professional sports science laboratories; today, it has become feasible in everyday life. The core of this technology lies in ‘pose estimation algorithms,’ such as OpenPose, MediaPipe, and DeepLabCut. These algorithms can recognize skeletal keypoint positions and further derive variables such as velocity, angular changes, and coordination. For example, by analyzing the rate of angular change in knee and hip joints, one can determine whether compensatory movements occur during squats, or assess stability when going up and down stairs.
RkJQdWJsaXNoZXIy MjIwMjA1