Mapping and Understanding Animal Behaviors with an Unsupervised 3D Motion Learning Framework

This presentation introduces a hierarchical 3D-motion learning framework for mapping spontaneous animal behavior and the related products, BehaviorAtlas. This technology integrates multi-view stereoscopic video capture and deep learning-based pose estimation to reconstruct full 3D pose skeleton. This framework transforms dynamic 3D trajectories into quantifiable behavior sequences, enabling comprehensive and precise behavioral analysis across various species and scenarios. BehaviorAtlas combines hardware and software to offer an integrated, user-friendly tool for researchers, with significant applications in neuroscience, pharmacology, and AI-driven behavioral studies.

Talks

• Dr. Kang Huang (Shenzhen Bayone Biotech)
  Shenzhen, China

Organisator*innen