1The University of Osaka 2Kyoto University 3Tokyo Institute of Technology
4National Institute of Advanced Industrial Science and Technology (AIST) 5Visual Geometry Group, University of Oxford
Example images of AnimalClue. We present AnimalClue, a dataset designed for identifying animal species based on their traces. Our dataset includes footprints, feces, eggs, bones, and feathers, totaling 159,605 bounding boxes from 968 animal species. We also annotate 22 traits such as habitat, diet, and activity pattern. We establish four benchmarks for evaluating models: classification, detection, instance segmentation, and traits classification.
Wildlife observation plays an important role in biodiversity conservation, necessitating robust methodologies for monitoring wildlife populations and interspecies interactions. Recent advances in computer vision have significantly contributed to automating fundamental wildlife observation tasks, such as animal detection and species identification. However, accurately identifying species from indirect evidence like footprints and feces remains relatively underexplored, despite its importance in contributing to wildlife monitoring. To bridge this gap, we introduce AnimalClue, the first large-scale dataset for species identification from images of indirect evidence. Our dataset consists of 159,605 bounding boxes encompassing five categories of indirect clues: footprints, feces, eggs, bones, and feathers. It covers 968 species, 200 families, and 65 orders. Each image is annotated with species-level labels, bounding boxes or segmentation masks, and fine-grained trait information, including activity patterns and habitat preferences. Unlike existing datasets primarily focused on direct visual features (e.g., animal appearances), AnimalClue presents unique challenges for classification, detection, and instance segmentation tasks due to the need for recognizing more detailed and subtle visual features. In our experiments, we extensively evaluate representative vision models and identify key challenges in animal identification from their traces.
Our dataset consists of 968 species, 200 families, and 65 orders. It includes a total of 159,605 bounding boxes across five trace types:
The total number of bounding boxes matches that of the classification dataset, while the number of images aligns with the detection and segmentation datasets.
While direct animal identification has been extensively studied, there remains significant potential to explore indirect methods—such as identifying animals through the traces they leave behind. Our AnimalClue dataset comprises five trace types and 968 species, with a total of 159,605 bounding boxes. It supports a wide range of tasks and includes fine-grained annotations.
Comparison with previous animal tracking datasets. CLS, DET, and SEG indicate classification, detection, and instance segmentation, respectively. Our AnimalClue contains diverse species and more number of bounding boxes.
Classification accuracy for all, frequent, and rare categories of animal specie.Throughout the species, family, and order categorization, Swin-B model tends to be higher accuracies on AnimalClue.
Visualization of t-SNE. By using a labeled dataset specialized for observing indirect animal clues, the separability among categories have been improved. When visualized in the feature space, the categories are better distinct.
@article{shinoda2025animalcluerecognizinganimalstraces, title={AnimalClue: Recognizing Animals by their Traces}, author={Risa Shinoda and Nakamasa Inoue and Iro Laina and Christian Rupprecht and Hirokatsu Kataoka}, year={2025}, eprint={2507.20240}, archivePrefix={arXiv}, primaryClass={cs.CV}, url={https://arxiv.org/abs/2507.20240}, }