Markerless Tracking

Markerless tracking uses computer vision and machine learning to estimate the pose (position and orientation) of objects or humans without the need for physical markers, aiming to automate data acquisition in various fields. Current research focuses on improving accuracy and robustness of pose estimation, particularly using deep learning models like MoveNet, ViTPose, and DeepLabCut, often incorporating techniques like multi-camera views and graph-based approaches for improved 3D tracking and handling of occlusions. This technology offers significant advantages in applications ranging from biomechanics analysis and gait assessment to robotics and augmented reality, enabling more accessible, cost-effective, and less intrusive data collection for a wide range of scientific and practical purposes.