19/01/2022 - Tommaso (ESR #6)
Today, we talk a little bit about tracking systems! Tracking systems are incredibly useful and powerful when it comes to analyzing animal behavior based on observations.Just try to imagine yourself busy watching hours and hours and hours of videos with animals freely moving in a determined space and having to try to understand the behavior exhibited. Not only it would require a huge effort, but it would also be very prone to mistakes if the person in charge has to do everything by him/herself.There are different approaches to achieve good tracking. Generally speaking, we can say that they all require a combination of hardware and software.The images you are obtained with one of the most popular system of tracking called DeepLabCut [1][2]. DeepLabCut is a a software that uses neural network processes to train algorithms and uses these to track the animals in the videos. A great advantage of this system is that require only a camera, the DeepLabCut software (which is open-source), and a little will to learn how it works. Check out the references to learn how can you use it!References:
- Mathis et al. 2018, DeepLabCut: markerless pose estimation of user-defined body parts with deep learning, Nature Neuroscience, https://doi.org/10.1038/s41593-018-0209-y
- Nath et al. 2019, Using DeepLabCut for 3D markerless pose estimation across species and behaviors, Nature protocols, https://doi.org/10.1038/s41596-019-0176-0