Overview

Note

This website is a companion for Talmo's lecture at the CAJAL QABVR 2024 course.         Learn more about SLEAP

Are you ready to get some SLEAP?   Let's get started!

Tutorial

1. Setup: Request a VM with SLEAP preinstalled (recommended) or install SLEAP locally.
2. Importing data: Setup a project and get your videos into SLEAP.
3. Initial labeling: Learn the interface and annotate your first frames.
4. Training a model: Configure and train your first SLEAP model.
5. Correcting predictions: Improve the model by correcting predictions.
6. Tracking new data: Apply your trained model to automate tracking on new videos.
7. Proofreading: Visualize the results and quickly fix tracking errors.
8. Exporting the results: Convert SLEAP predictions to other formats.
9. I'm done SLEAPing, now what?: Analyze your pose trajectories yourself or with other tools.

Useful links

• SLEAP Docs
• SLEAP Docs → Tutorial
• SLEAP GitHub
• sleap-io: Standalone library for I/O with pose data.
• Talmo Lab

Acknowledgments

Big thanks to Andrew Park, Ramiz Hajj and Liezl Maree for putting together the cloud platform that supports this tutorial! We also acknowledge the brave TAs who helped lead the first iteration of this tutorial at Cosyne 2024: Shayan Afshar, Elizabeth Berrigan, Jianna Cressy, Kosmas Deligkaris, Chaitanya Kapoor, Chris Lee, Eric Leonardis, Aaditya Prasad.