DeepPoseKit, a software toolkit for fast and robust animal pose estimation using deep learning
- Max Planck Institute of Animal Behavior, Germany
- Princeton University, United States
Abstract
Quantitative behavioral measurements are important for answering questions across scientific disciplines-from neuroscience to ecology. State-of-the-art deep-learning methods offer major advances in data quality and detail by allowing researchers to automatically estimate locations of an animal's body parts directly from images or videos. However, currently-available animal pose estimation methods have limitations in speed and robustness. Here we introduce a new easy-to-use software toolkit, DeepPoseKit, that addresses these problems using an efficient multi-scale deep-learning model, called Stacked DenseNet, and a fast GPU-based peak-detection algorithm for estimating keypoint locations with subpixel precision. These advances improve processing speed >2× with no loss in accuracy compared to currently-available methods. We demonstrate the versatility of our methods with multiple challenging animal pose estimation tasks in laboratory and field settings-including groups of interacting individuals. Our work reduces barriers to using advanced tools for measuring behavior and has broad applicability across the behavioral sciences.
Data availability
Data used and generated for experiments and model comparisons are included in the supporting files. Posture datasets can be found at: https://github.com/jgraving/deepposekit-dataThe code for DeepPoseKit is publicly available at the URL we provided in the paper: https://github.com/jgraving/deepposekit/The reviewers should follow the provided instructions for installation in the README file https://github.com/jgraving/deepposekit/blob/master/README.md#installation. Example Jupyter notebooks for how to use the code are provided here: https://github.com/jgraving/deepposekit/tree/master/examples
-
Fast animal pose estimation using deep neural networkshttp://arks.princeton.edu/ark:/88435/dsp01pz50gz79z.
Article and author information
Author details
Funding
National Science Foundation (IOS-1355061)
- Iain D Couzin
Horizon 2020 Framework Programme (Marie Sklodowska-Curie grant agreement No. 748549)
- Blair R Costelloe
Nvidia (GPU Grant)
- Blair R Costelloe
Office of Naval Research (N00014-09-1-1074)
- Iain D Couzin
Office of Naval Research (N00014-14-1-0635)
- Iain D Couzin
Army Research Office (W911NG-11-1-0385)
- Iain D Couzin
Army Research Office (W911NF14-1-0431)
- Iain D Couzin
Deutsche Forschungsgemeinschaft (DFG Centre of Excellence 2117)
- Iain D Couzin
University of Konstanz (Zukunftskolleg Investment Grant)
- Blair R Costelloe
The Strukture-und Innovations fonds fur die Forschung of the State of Baden-Wurttemberg
- Iain D Couzin
Max Planck Society
- Iain D Couzin
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
Ethics
Animal experimentation: All procedures for collecting the zebra (E. grevyi) dataset were reviewed and approved by Ethikrat, the independent Ethics Council of the Max Planck Society. The zebra dataset was collected with the permission of Kenya's National Commission for Science, Technology and Innovation (NACOSTI/P/17/59088/15489 and NACOSTI/P/18/59088/21567) using drones operated by B.R.C. with the permission of the Kenya Civil Aviation Authority (authorization numbers: KCAA/OPS/2117/4 Vol. 2 (80), KCAA/OPS/2117/4 Vol. 2 (81), KCAA/OPS/2117/5 (86) and KCAA/OPS/2117/5 (87); RPAS Operator Certificate numbers: RPA/TP/0005 AND RPA/TP/000-0009).
Copyright
© 2019, Graving et al.
This article is distributed under the terms of the Creative Commons Attribution License permitting unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 25,167
- views
-
- 2,447
- downloads
-
- 414
- citations
Views, downloads and citations are aggregated across all versions of this paper published by eLife.
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
DeepPoseKit, a software toolkit for fast and robust animal pose estimation using deep learning
eLife 8:e47994.
https://doi.org/10.7554/eLife.47994
Further reading
-
- Neuroscience
Detecting causal relations structures our perception of events in the world. Here, we determined for visual interactions whether generalized (i.e. feature-invariant) or specialized (i.e. feature-selective) visual routines underlie the perception of causality. To this end, we applied a visual adaptation protocol to assess the adaptability of specific features in classical launching events of simple geometric shapes. We asked observers to report whether they observed a launch or a pass in ambiguous test events (i.e. the overlap between two discs varied from trial to trial). After prolonged exposure to causal launch events (the adaptor) defined by a particular set of features (i.e. a particular motion direction, motion speed, or feature conjunction), observers were less likely to see causal launches in subsequent ambiguous test events than before adaptation. Crucially, adaptation was contingent on the causal impression in launches as demonstrated by a lack of adaptation in non-causal control events. We assessed whether this negative aftereffect transfers to test events with a new set of feature values that were not presented during adaptation. Processing in specialized (as opposed to generalized) visual routines predicts that the transfer of visual adaptation depends on the feature similarity of the adaptor and the test event. We show that the negative aftereffects do not transfer to unadapted launch directions but do transfer to launch events of different speeds. Finally, we used colored discs to assign distinct feature-based identities to the launching and the launched stimulus. We found that the adaptation transferred across colors if the test event had the same motion direction as the adaptor. In summary, visual adaptation allowed us to carve out a visual feature space underlying the perception of causality and revealed specialized visual routines that are tuned to a launch’s motion direction.
-
- Neuroscience
Synchronous neuronal activity is organized into neuronal oscillations with various frequency and time domains across different brain areas and brain states. For example, hippocampal theta, gamma, and sharp wave oscillations are critical for memory formation and communication between hippocampal subareas and the cortex. In this study, we investigated the neuronal activity of the dentate gyrus (DG) with optical imaging tools during sleep-wake cycles in mice. We found that the activity of major glutamatergic cell populations in the DG is organized into infraslow oscillations (0.01–0.03 Hz) during NREM sleep. Although the DG is considered a sparsely active network during wakefulness, we found that 50% of granule cells and about 25% of mossy cells exhibit increased activity during NREM sleep, compared to that during wakefulness. Further experiments revealed that the infraslow oscillation in the DG was correlated with rhythmic serotonin release during sleep, which oscillates at the same frequency but in an opposite phase. Genetic manipulation of 5-HT receptors revealed that this neuromodulatory regulation is mediated by Htr1a receptors and the knockdown of these receptors leads to memory impairment. Together, our results provide novel mechanistic insights into how the 5-HT system can influence hippocampal activity patterns during sleep.
