Social LEAP Estimates Animal Poses (SLEAP)
Type: Software,
Keywords: Academic software, Deep learning Software, Neural network, Anatomical landmark, BRAIN Initiative
Deep learning framework for estimating animal pose, Deep learning software framework for general purpose multi-animal limb tracking from video
SLEAP (Social LEAP Estimates Animal Poses) is a deep learning software framework for general purpose multi-animal limb tracking from video. This software couples a GUI for importing and annotating data with deep neural networks designed for learning to locate and associate user-specified anatomical landmarks on unmarked animals. Use cases range from kinematic studies of animal movement, to quantification of social dynamics via multi-animal part tracking.
* AI-driven multi-animal pose estimation software
* Open source software package. *Python-based
* Tracks individual body part kinematics of single or multiple interacting animals
* Supports multi-animal pose estimation, animal instance tracking
* Labeling/training GUI supports active learning
* Includes a graphical proof-reading tool to quickly assess the accuracy of tracking and correcting problems
* Framework for multi-animal body part position estimation via deep learning
* Tracking of individual body part kinematics of multiple animals
* High throughput analysis of animal behaviors (e.g., drug screens)
* Development of disease diagnostics for animal models
* Wildlife conservation and monitoring of animal well-being (e.g., agrotech)
* https://www.sciencedirect.com/science/article/pii/S2211124722009354
* https://plantmethods.biomedcentral.com/articles/10.1186/s13007-022-00851-9
* https://www.biorxiv.org/content/10.1101/2022.07.18.500505v1.abstract
* https://www.biorxiv.org/content/10.1101/2022.05.25.493496v1.abstract
* https://www.nature.com/articles/s41586-021-03962-w
* https://link.springer.com/article/10.1186/s13229-022-00492-8
* Mice, flies, bees (compatible with any animal)
* Learns from few user examples
* Fast training of deep neural networks on consumer hardware
* Highly accurate and robust to variability in imaging conditions
* Single or multiple animal tracking
* No programming skills required
* For research/academic use only
* Compatible with Python versions 3.6 and above, with support for Windows and Linux
* Mac OS X works but without GPU support
* Remote cloud/cluster-based training and inference
* Pereira et al. 2018, Fast animal pose estimation using deep neural networks, Nature Methods 16: 117–125
* https://www.nature.com/articles/s41592-022-01426-1
* https://sleap.ai
* https://pypi.org/project/sleap/
* https://www.nature.com/articles/s41592-018-0234-5
* https://www.biorxiv.org/content/10.1101/2020.08.31.276246v1.abstract
* Recommended to use the Anaconda Python distribution for Windows
* Pip package available for all systems
Talmo Pereira (Salk Fellow & PI)
Mala Murthy (Professor)
Joshua Shaevitz (Professor)
Princeton
TEAM / COLLABORATOR(S)
Talmo Pereira, Salk Institute for Biological Studies
Liezl Maree, Salk Institute for Biological Studies
Arlo Sheridan, Salk Institute for Biological Studies
David Samy, Salk Institute for Biological Studies
Arie Matsliah, Princeton Neuroscience Institute, Princeton University
Nat Tabris, Princeton Neuroscience Institute, Princeton University
David Turner, Research Computing and Princeton Neuroscience Institute, Princeton University
Joshua Shaevitz, Physics and Lewis-Sigler Institute, Princeton University
Mala Murthy, Princeton Neuroscience Institute, Princeton University
FUNDING SOURCE(S)
NIH BRAIN Initative R01 NS104899
Princeton Innovation Accelerator Fund (NSF GRFP, NIH R01, HHMI)