Before you can use SLEAP, you’ll need to install it. Follow the instructions at Installation to install SLEAP and start the GUI app.
There are three main stages of using SLEAP:
Creating a project, opening a movie and defining the skeleton;
Labeling and learning, labeling of video frames assisted by network predictions;
Prediction and proofreading, final network predictions of body-part positions and proofreading of track identities in full videos.
The tutorial will walk you through this entire process. You can read through the tutorial and then try running SLEAP on our sample datasets—this will allow you to try out training and inference before you’ve labeled your own data for training.
- Creating a project
- Initial Labeling
- Quick Training and Inference
- Prediction-assisted labeling
- Reviewing and fixing predictions
- Tracking instances across frames
- Export Data For Analysis