DeepWild contains all our code, protocols, and data on our DeepLabCut trained full body wild ape models
twitter.example.mp4
DeepWild is a machine learning model trained to track the movement of great apes in video data, created using DeepLabCut The model tracks 18 key points along an apes body. It can track multiple individuals and has been trained to work with wild data.
You can download the most recent version of the model on our google drive Link: https://drive.google.com/drive/folders/1lGF3vaX2Qz8nasnS9NAZ1NKkI8DkHXrr We currently have two models available. Deepwild1.0: This is the model demonstrated in our paper. You can use the model 'off the shelf'- input your video data into the model and it will track your apes. Or, you can use this model as a base to change the model to suit your needs. You may want to add new keypoints or add new videos to the training dataset. Deepwild1.1: This is our most recent version of the model, containing more training frames from more populations. You can use the model 'off the shelf'- input your video data into the model and it will track your apes. If you would like to use this model as a base to change the model to suit your needs, please contact us. Some of the frames and videos needed to re-train the model are not open source.
DeepWild: Application of the pose estimation tool DeepLabCut for behaviour tracking in wild chimpanzees and bonobos By: Charlotte Wiltshire, James Lewis-Cheetham, Viola Komedová, Tetsuro Matsuzawa, Kirsty E. Graham, Catherine Hobaiter Link to paper: https://besjournals.onlinelibrary.wiley.com/doi/full/10.1111/1365-2656.13932
If you would like to contribute to DeepWild, you can label frames to help train the model here: https://contrib.deeplabcut.org/label It takes a few seconds to mark a point, give it a go!
