Comparison of a single-view image-based system to a multi-camera marker-based system for human static pose estimation

Evaluation of 3D Markerless Motion Capture System Accuracy during Skate Skiing on a Treadmill

In this study, we developed a deep learning-based 3D markerless motion capture system for skate skiing on a treadmill and evaluated its accuracy against marker-based motion capture during G1 and G3 skating techniques. Participants performed roller skiing trials on a skiing treadmill. Trials were recorded with two synchronized video cameras (100 Hz). We then trained a custom model using DeepLabCut, and the skiing movements were analyzed using both DeepLabCut-based markerless motion capture and marker-based motion capture systems. We statistically compared joint centers and joint vector angles between the methods. The results demonstrated a high level of agreement for joint vector angles, with mean differences ranging from -2.47° to 3.69°. For joint center positions and toe placements, mean differences ranged from 24.0 to 40.8 mm. This level of accuracy suggests that our markerless approach could be useful as a skiing coaching tool. The method presents interesting opportunities for capturing and extracting value from large amounts of data without the need for markers attached to the skier and expensive cameras.

Feasibility of 3D Body Tracking from Monocular 2D Video Feeds in Musculoskeletal Telerehabilitation

Musculoskeletal conditions affect millions of people globally; however, conventional treatments pose challenges concerning price, accessibility, and convenience. Many telerehabilitation solutions offer an engaging alternative but rely on complex hardware for body tracking. This work explores the feasibility of a model for 3D Human Pose Estimation (HPE) from monocular 2D videos (MediaPipe Pose) in a physiotherapy context, by comparing its performance to ground truth measurements. MediaPipe Pose was investigated in eight exercises typically performed in musculoskeletal physiotherapy sessions, where the Range of Motion (ROM) of the human joints was the evaluated parameter. This model showed the best performance for shoulder abduction, shoulder press, elbow flexion, and squat exercises. Results have shown a MAPE ranging between 14.9% and 25.0%, Pearson's coefficient ranging between 0.963 and 0.996, and cosine similarity ranging between 0.987 and 0.999. Some exercises (e.g., seated knee extension and shoulder flexion) posed challenges due to unusual poses, occlusions, and depth ambiguities, possibly related to a lack of training data. This study demonstrates the potential of HPE from monocular 2D videos, as a markerless, affordable, and accessible solution for musculoskeletal telerehabilitation approaches. Future work should focus on exploring variations of the 3D HPE models trained on physiotherapy-related datasets, such as the Fit3D dataset, and post-preprocessing techniques to enhance the model's performance.