A review of 28 free animal-tracking software applications: current features and limitations

Veronica Panadeiro,Alvaro Rodriguez,Jason Henry,Donald Wlodkowic,Magnus Andersson
DOI: https://doi.org/10.1038/s41684-021-00811-1
2021-07-29
Lab Animal
Abstract:Well-quantified laboratory studies can provide a fundamental understanding of animal behavior in ecology, ethology and ecotoxicology research. These types of studies require observation and tracking of each animal in well-controlled and defined arenas, often for long timescales. Thus, these experiments produce long time series and a vast amount of data that require the use of software applications to automate the analysis and reduce manual annotation. In this review, we examine 28 free software applications for animal tracking to guide researchers in selecting the software that might best suit a particular experiment. We also review the algorithms in the tracking pipeline of the applications, explain how specific techniques can fit different experiments, and finally, expose each approach’s weaknesses and strengths. Our in-depth review includes last update, type of platform, user-friendliness, off- or online video acquisition, calibration method, background subtraction and segmentation method, species, multiple arenas, multiple animals, identity preservation, manual identity correction, data analysis and extra features. We found, for example, that out of 28 programs, only 3 include a calibration algorithm to reduce image distortion and perspective problems that affect accuracy and can result in substantial errors when analyzing trajectories and extracting mobility or explored distance. In addition, only 4 programs can directly export in-depth tracking and analysis metrics, only 5 are suited for tracking multiple unmarked animals for more than a few seconds and only 11 have been updated in the period 2019–2021.
veterinary sciences
What problem does this paper attempt to address?