AthleticsPose: Authentic Sports Motion Dataset on Athletic Field and Evaluation of Monocular 3D Pose Estimation Ability

2025-07-17

AI Summary

Monocular 3D human pose estimation holds significant promise for sports analytics, yet its practical deployment is hindered by the scarcity of real-world athletic motion data and unclear reliability in field conditions. To address this, we introduce the first benchmark tailored to realistic outdoor track-and-field scenarios—AthleticsPose—a high-fidelity dataset featuring 23 athletes, multi-view synchronized recordings, and high-speed motion sequences. This fills a critical gap in high-quality monocular 3D pose benchmarks for sports. Rigorous evaluation using MPJPE and joint-angle analysis reveals severe domain gaps: models trained on synthetic or imitation data generalize poorly to real athletics. In contrast, real-data training reduces MPJPE by ~75% and markedly improves modeling of inter-subject knee-angle variability. However, temporal modeling errors persist during high-speed motion. Our work establishes a reliable data foundation and methodological framework for sports motion analysis.

Technology Category

• Computer Vision: Motion & Tracking
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Learning with Manifolds

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Monocular 3D pose estimation is a promising, flexible alternative to costly motion capture systems for sports analysis. However, its practical application is hindered by two factors: a lack of realistic sports datasets and unclear reliability for sports tasks. To address these challenges, we introduce the AthleticsPose dataset, a new public dataset featuring ``real'' motions captured from 23 athletes performing various athletics events on an athletic field. Using this dataset, we trained a representative 3D pose estimation model and performed a comprehensive evaluation. Our results show that the model trained on AthleticsPose significantly outperforms a baseline model trained on an imitated sports motion dataset, reducing MPJPE by approximately 75 %. These results show the importance of training on authentic sports motion data, as models based on imitated motions do not effectively transfer to real-world motions. Further analysis reveals that estimation accuracy is sensitive to camera view and subject scale. In case studies of kinematic indicators, the model demonstrated the potential to capture individual differences in knee angles but struggled with higher-speed metrics, such as knee-drive velocity, due to prediction biases. This work provides the research community with a valuable dataset and clarifies the potential and practical limitations of using monocular 3D pose estimation for sports motion analysis. Our dataset, code, and checkpoints are available at https://github.com/SZucchini/AthleticsPose.

Problem

• Lack of realistic sports datasets for 3D pose estimation
• Unclear reliability of monocular 3D pose estimation in sports
• Sensitivity of estimation accuracy to camera view and scale

Innovation

• Introduces AthleticsPose dataset for real sports motions
• Trains 3D pose model reducing MPJPE by 75%
• Evaluates accuracy sensitivity to camera view