A MATLAB® community toolbox for applying the SLEAP multi-animal pose estimation deep learning framework [1]. This toolbox is intended to make it easy to use SLEAP models natively in MATLAB.
🚧 SLEAP Toolbox is early stage. Interested in using it or helping to create future versions? See contact info below.
👀 See SLEAP Toolbox in action with this live script demo.
SLEAP is an open source deep-learning based framework for multi-animal pose tracking. It can be used to track any type or number of animals by training neural networks from user-labeled frames.
Currently, SLEAP Toolbox supports inference based on top-down models trained in SLEAP. To use it, you must first train a set of models from the SLEAP GUI (see tutorial).
SLEAP Toolbox supports the following key steps:
- Import of top-down model components from SLEAP into MATLAB
- Composition of a single composite top-down model as a
DagNetworkobject - Prediction of animal pose estimates per frame of user data
- Visualization of animal pose estimates for user data within the MATLAB graphics system
The prediction step is carried out simply, via the function predict in the Deep Learning Toolbox™.
Pretrained models and sample data are included to help get started.
Usage of SLEAP Toolbox is subject to its license terms.
Installation via the Add-on Explorer is recommended. This will install the latest version from GitHub, adding the SLEAP Toolbox root folder to your MATLAB path.
Note: It is NOT necessary to have SLEAP installed in order to run this toolbox. You can train SLEAP on another machine, and then use this toolbox to predict on new data within MATLAB.
See the demo.
For support for this toolbox, please open a GitHub issue in this repository.
For general SLEAP support, check out the Official Documentation or the support forums on GitHub.
Please direct other inquiries and interest(❕) to [email protected] with "SLEAP Toolbox" included in the subject line.
[1] T.D. Pereira, N. Tabris, A. Matsliah, D. M. Turner, J. Li, S. Ravindranath, E. S. Papadoyannis, E. Normand, D. S. Deutsch, Z. Y. Wang, G. C. McKenzie-Smith, C. C. Mitelut, M. D. Castro, J. D’Uva, M. Kislin, D. H. Sanes, S. D. Kocher, S. S-H, A. L. Falkner, J. W. Shaevitz, and M. Murthy. SLEAP: A deep learning system for multi-animal pose tracking. Nature Methods, pp.1-10, 2022
Copyright (c) 2019 - 2022 As Indicated Per File.
Distributed under license permitting academic and research use.
For commercial inquiries on using SLEAP for commercial applications, please see the relevant section in the main SLEAP repository.