From Sports Videos to Immersive Training: Augmenting Human Motion to Enrich Basketball Training E...
summary
Pipeline reconstructs 3D player and ball motion from broadcast video using computer-vision tools (object detection, trajectory fitting, 4D Humans, EasyMocap), turns shots into an MR one-on-one with fouls and jump-triggered defender reactions, and exposes KPIs like arm swing, release angle, and release height. Two studies (pairs defending three fake/three real shots; 12 shooters — 6 high, 6 amateur) show improved defensive timing and intuition, but MR pressure lowers accuracy for skilled players.
AI-written summary · about
chapters (9)
- 0:00Introduction and overview
- 0:45Challenges with sports video training
- 2:53Research goals and framework
- 3:57Motion data extraction process
- 4:40Immersive visualization design
- 5:22One-on-one interactive training
- 6:46User study results and findings
- 8:53Limitations and future work
- 9:19Q&A session
related talks
43mBeyond The Printed Page
06mCoolMoves: User Motion Accentuation in Virtual Reality
48mTalking VR Turkey
20mInteractive Feedforward for Improving Performance and Maintaining Intrinsic Motivation in VR ...
41mComputer Vision for Remote AR
14mHow Brands are Championing the AR Media Revolution
34mTransforming Education with XR and AI
32mNext generation human character control