Abstract
Quantifying movement is critical for understanding animal behavior. Advances in computer vision now enable markerless tracking from 2D video, but most animals live and move in 3D. Here, we introduce Anipose, a Python toolkit for robust markerless 3D pose estimation. Anipose is built on the popular 2D tracking method DeepLabCut, so users can easily expand their existing experimental setups to obtain accurate 3D tracking. It consists of four components: (1) a 3D calibration module, (2) filters to resolve 2D tracking errors, (3) a triangulation module that integrates temporal and spatial regularization, and (4) a pipeline to structure processing of large numbers of videos. We evaluate Anipose on four datasets: a moving calibration board, fruit flies walking on a treadmill, mice reaching for a pellet, and humans performing various actions. By analyzing 3D leg kinematics tracked with Anipose, we identify a key role for joint rotation in motor control of fly walking. We believe this open-source software and accompanying tutorials (anipose.org) will facilitate the analysis of 3D animal behavior and the biology that underlies it.
Competing Interest Statement
The authors have declared no competing interest.
Footnotes
↵* Co-senior authors: bbrunton{at}uw.edu, tuthill{at}uw.edu
Major revisions to the text and additional figures