WildlifeReID-10k: Wildlife re-identification dataset with 10k individual animals
This provides a standardized benchmark for researchers working on multi-species animal re-identification, though it is incremental as it re-samples existing datasets.
The paper tackles the problem of wildlife re-identification by introducing WildlifeReID-10k, a large-scale dataset with over 10k animal identities and 140k images, which challenges state-of-the-art methods and includes protocols to prevent data leakage.
This paper introduces WildlifeReID-10k, a new large-scale re-identification benchmark with more than 10k animal identities of around 33 species across more than 140k images, re-sampled from 37 existing datasets. WildlifeReID-10k covers diverse animal species and poses significant challenges for SoTA methods, ensuring fair and robust evaluation through its time-aware and similarity-aware split protocol. The latter is designed to address the common issue of training-to-test data leakage caused by visually similar images appearing in both training and test sets. The WildlifeReID-10k dataset and benchmark are publicly available on Kaggle, along with strong baselines for both closed-set and open-set evaluation, enabling fair, transparent, and standardized evaluation of not just multi-species animal re-identification models.