# ZebraPose: Zebra Detection and Pose Estimation using only Synthetic Data

arXiv: https://arxiv.org/abs/2408.10831 · HTML: https://mlsanity.com/paper/2408.10831/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 43 references that are in the map.
- First posted on arXiv 20 Aug 2024.

## Builds on (most cited first, 22 of 22)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://mlsanity.com/paper/2303.05499/) (2023)
- [AMASS: Archive of Motion Capture as Surface Shapes](https://mlsanity.com/paper/1904.03278/) (2019)
- [3D Menagerie: Modeling the 3D shape and pose of animals](https://mlsanity.com/paper/1611.07700/) (2016)
- [BEDLAM: A Synthetic Dataset of Bodies Exhibiting Detailed Lifelike Animated Motion](https://mlsanity.com/paper/2306.16940/) (2023)
- [AP-10K: A Benchmark for Animal Pose Estimation in the Wild](https://mlsanity.com/paper/2108.12617/) (2021)
- [Cross-Domain Adaptation for Animal Pose Estimation](https://mlsanity.com/paper/1908.05806/) (2019)
- [Three-D Safari: Learning to Estimate Zebra Pose, Shape, and Texture from Images "In the Wild"](https://mlsanity.com/paper/1908.07201/) (2019)
- [Learning from Synthetic Animals](https://mlsanity.com/paper/1912.08265/) (2019)
- [ViTPose++: Vision Transformer for Generic Body Pose Estimation](https://mlsanity.com/paper/2212.04246/) (2022)
- [APT-36K: A Large-scale Benchmark for Animal Pose Estimation and Tracking](https://mlsanity.com/paper/2206.05683/) (2022)
- [From Synthetic to Real: Unsupervised Domain Adaptation for Animal Pose Estimation](https://mlsanity.com/paper/2103.14843/) (2021)
- [Synthetic Examples Improve Generalization for Rare Classes](https://mlsanity.com/paper/1904.05916/) (2019)
- [CLAMP: Prompt-based Contrastive Learning for Connecting Language and Animal Pose](https://mlsanity.com/paper/2206.11752/) (2022)
- [Re-Thinking Inverse Graphics With Large Language Models](https://mlsanity.com/paper/2404.15228/) (2024)
- [PSP-HDRI+: A Synthetic Dataset Generator for Pre-Training of Human-Centric Computer Vision Models](https://mlsanity.com/paper/2207.05025/) (2022)
- [ScarceNet: Animal Pose Estimation with Scarce Annotations](https://mlsanity.com/paper/2303.15023/) (2023)
- [AirPose: Multi-View Fusion Network for Aerial 3D Human Pose and Shape Estimation](https://mlsanity.com/paper/2201.08093/) (2022)
- [Prior-Aware Synthetic Data to the Rescue: Animal Pose Estimation with Very Limited Real Data](https://mlsanity.com/paper/2208.13944/) (2022)
- [Synthetic Data-based Detection of Zebras in Drone Imagery](https://mlsanity.com/paper/2305.00432/) (2023)
- [SPAC-Net: Synthetic Pose-aware Animal ControlNet for Enhanced Pose Estimation](https://mlsanity.com/paper/2305.17845/) (2023)
- [Behavior Discovery and Alignment of Articulated Object Classes from Unstructured Video](https://mlsanity.com/paper/1511.09319/) (2015)

## Built on by (most cited first, 2 of 2)

- [MOO: A Multi-view Oriented Observations Dataset for Viewpoint Analysis in Cattle Re-Identification](https://mlsanity.com/paper/2603.04314/) (2026)
- [Promptable Animal Pose Tracking Across Species](https://mlsanity.com/paper/2608.04995/) (2026)

## Abstract (arXiv)

Collecting and labeling large real-world wild animal datasets is impractical, costly, error-prone, and labor-intensive. For animal monitoring tasks, as detection, tracking, and pose estimation, out-of-distribution viewpoints (e.g. aerial) are also typically needed but rarely found in publicly available datasets. To solve this, existing approaches synthesize data with simplistic techniques that then necessitate strategies to bridge the synthetic-to-real gap. Therefore, real images, style constraints, complex animal models, or pre-trained networks are often leveraged. In contrast, we generate a fully synthetic dataset using a 3D photorealistic simulator and demonstrate that it can eliminate such needs for detecting and estimating 2D poses of wild zebras. Moreover, existing top-down 2D pose estimation approaches using synthetic data assume reliable detection models. However, these often fail in out-of-distribution scenarios, e.g. those that include wildlife or aerial imagery. Our method overcomes this by enabling the training of both tasks using the same synthetic dataset. Through extensive benchmarks, we show that models trained from scratch exclusively on our synthetic data generalize well to real images. We perform these using multiple real-world and synthetic datasets, pre-trained and randomly initialized backbones, and different image resolutions. Code, results, models, and data can be found athttps://zebrapose.is.tue.mpg.de/.