Mocap as a Service: Video Motion Capture Makes Human Motion Analysis for Everyone

Mocap as a Service: Video Motion Capture Makes Human Motion Analysis for Everyone

by Yoshihiko Nakamura

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQVComputer vision
🎙 Yoshihiko Nakamura 👥 2K 📅 May 29, 2019 ⏱ 40 min 👁 588 📄 expert opinion 🧭 2026-08-18
Available in: English (current) Français

Keywords

motion capturehuman motion analysismuscle activityvideo-based mocapcloud service

Summary

In this plenary talk, Professor Yoshihiko Nakamura from the University of Tokyo presents his vision of making motion capture technology accessible to everyone through video-based systems and cloud computing. He begins by introducing his lab’s work on humanoid robotics, emphasizing the importance of understanding human motion for developing natural robot behaviors. He then describes traditional marker-based and IMU-based motion capture systems, highlighting their limitations in terms of setup time and cost. The core of the talk focuses on their recent development of a video-based motion capture system that uses multiple cameras and deep learning (specifically OpenPose) to reconstruct 3D skeletal movements without markers. This system is demonstrated in various applications, including analyzing the muscle activities of elite athletes such as Olympic medalists and professional football players. The ultimate goal is to provide ‘Mocap as a Service’ over high-speed academic networks, enabling widespread use in sports training, healthcare, and other fields. The talk concludes with a Q&A session addressing cost, integration with robotics, and potential for energy expenditure estimation.

168 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk provides valuable insights into the state-of-the-art in motion capture and its applications. The speaker demonstrates a clear progression from research to practical implementation, supported by concrete examples and collaborations. The argumentation is solid, grounded in years of research and real-world case studies, though some technical details are glossed over.

Scientific Rigor, Source Quality, Title Accuracy

The speaker is a highly respected researcher, and the content is based on original research and collaborations. However, specific citations are not provided in the talk, and the description lacks links to papers or resources. The title accurately reflects the content, and the talk is well-structured.

112 words

Title / Content Match

The title accurately reflects the content, focusing on making motion capture accessible via video and cloud services.

Quality & Reliability

8/10

The talk is delivered by a leading researcher in humanoid robotics and motion capture, with extensive experience and peer-reviewed publications. The content is based on original research and practical applications, but some claims lack detailed methodological transparency.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Contribution & Novelties

The talk presents a novel approach to making motion capture accessible via video and cloud services, potentially democratizing human motion analysis. The integration of deep learning-based pose estimation with biomechanical modeling is a significant contribution.

Pour aller plus loin :

  • OpenPose — The deep learning tool used for 2D pose estimation.
  • Musculoskeletal model — Background on modeling muscles and bones.
  • Inverse kinematics — Technique used to reconstruct joint angles from motion data.

72 words

Radar Profile

The radar profile shows high scores in information quantity, quality, and reliability, with a slightly lower technical level, indicating a comprehensive and credible presentation that is accessible to a broad audience.

Reliability 8/10