Social LEAP Estimates Animal Pose (SLEAP) is a framework for multi-animal body part position estimation via deep learning. It is the successor to LEAP. SLEAP is written entirely in Python, supports multi-animal pose estimation, animal instance tracking, and comes with a labeling/training GUI that supports active learning.
SLEAP is released under a Clear BSD License and is intended for research/academic use only. For commercial use, please contact: Laurie Tzodikov (Assistant Director, Office of Technology Licensing), Princeton University, 609-258-7256.
If you use SLEAP in your research, please cite:
Pereira, Talmo D., Nathaniel Tabris, Junyu Li, Shruthi Ravindranath, Eleni S. Papadoyannis, Z. Yan Wang, David M. Turner, et al. 2020. “SLEAP: Multi-Animal Pose Tracking.” bioRxiv. https://doi.org/10.1101/2020.08.31.276246.