October 9, 2025Open Access

Paving the Way Towards Kinematic Assessment Using Monocular Video: A Preclinical Benchmark of State-of-the-Art Deep-Learning-Based 3D Human Pose Estimators Against Inertial Sensors in Daily Living Activities

View Full Paper

Ask AI

Bookmark

Authors

MMMario Medrano-ParedesUniversidad de Valladolid CFCarmen Fernández-GonzálezUniversidad de Valladolid FPFrancisco Javier Díaz Pernas

Discussion

Member takes

Implication

Preclinical analysis compares video-based 3D human pose estimators to inertial sensors, suggesting both can improve telemedicine and rehabilitation outcomes.

Key Points

MotionAGFormer showed superior performance among models, achieving the lowest RMSE and MAE scores.
The study evaluated joint angles from monocular video against IMU data using OpenSim inverse kinematics.
Key trade-offs between video and sensor-based approaches were identified, including costs and precision.
Findings highlight the potential of off-the-shelf video models for clinically relevant kinematic assessments.

Cite This Study

Medrano-Paredes et al. (2025) studied this question.

synapsesocial.com/papers/68e7d631bd66d359be62669a https://doi.org/10.48550/arxiv.2510.02264

View Full Paper

Ask AI

Bookmark

Implication

Preclinical analysis compares video-based 3D human pose estimators to inertial sensors, suggesting both can improve telemedicine and rehabilitation outcomes.

Paving the Way Towards Kinematic Assessment Using Monocular Video: A Preclinical Benchmark of State-of-the-Art Deep-Learning-Based 3D Human Pose Estimators Against Inertial Sensors in Daily Living Activities

Discussion

Member takes

Implication

Key Points

Cite This Study

Paving the Way Towards Kinematic Assessment Using Monocular Video: A Preclinical Benchmark of State-of-the-Art Deep-Learning-Based 3D Human Pose Estimators Against Inertial Sensors in Daily Living Activities

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider