idtracker.ai: Tracking all individuals in large collectives of unmarked animals
Abstract
Our understanding of collective animal behavior is limited by our ability to track each of the individuals. We describe an algorithm and software, idtracker.ai, that extracts from video all trajectories with correct identities at a high accuracy for collectives of up to 100 individuals. It uses two deep networks, one detecting when animals touch or cross and another one for animal identification, trained adaptively to conditions and difficulty of the video.
- Publication:
-
arXiv e-prints
- Pub Date:
- March 2018
- DOI:
- 10.48550/arXiv.1803.04351
- arXiv:
- arXiv:1803.04351
- Bibcode:
- 2018arXiv180304351R
- Keywords:
-
- Computer Science - Computer Vision and Pattern Recognition
- E-Print:
- 44 pages, 1 main figure, 13 supplementary figures, 6 tables