IDLE-OO Camera Traps: Balanced Five-Set Species Classification Benchmark from LILA BC Camera Trap Images (Imageomics, BioCLIP 2)

Provider
Elizabeth Campolongo, Jianyang Gu, Net Zhang (Imageomics Institute)
License
CDLA-Permissive-1.0 (source)
Size
2,613 files
Format
CSV, PNG (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/imageomics/IDLE-OO-Camera-Traps)
Access
Free and public. No account with the data provider is needed.
DOI
10.57967/hf/10148
Category
Life Sciences
Source
https://huggingface.co/datasets/imageomics/IDLE-OO-Camera-Traps

About this dataset

A species classification benchmark of 2,586 camera trap images drawn from five datasets in the Labeled Information Library of Alexandria: Biology and Conservation (LILA BC), each balanced to the same number of images per species, built by the Imageomics Institute to evaluate the BioCLIP 2 biology foundation model in zero-shot, one-shot and five-shot settings. The five sets are Desert Lion Conservation Camera Traps from Namibia (352 images, 32 species), ENA24-detection from eastern North America (1,120 images, 20 species), Island Conservation Camera Traps (310 images, 33 common names), Ohio Small Animals (468 images) and Orinoquia Camera Traps from Colombia (336 images). Images sit under data/test/ / as PNG with a metadata.csv; per-set balanced CSVs give scientific and common names and full seven-rank taxonomy, and potential-sets/ holds the filtered LILA image URL and label tables plus the notebooks that sampled them. Suited to evaluating CLIP-style models (CLIP, SigLIP, BioCLIP, BioCLIP 2) on wildlife camera trap images. License CDLA-Permissive-1.0, the same as the LILA source images. Cite the dataset (Campolongo, Gu, Zhang 2026, DOI 10.57967/hf/10148), BioCLIP 2 and each source dataset.

How to load this dataset in Python

import pandas as pd
df = pd.read_csv("hf://datasets/imageomics/IDLE-OO-Camera-Traps/ENA24-balanced.csv")
df.head()

Data structure

File types in sampled files: CSV, PNG

Files

Example file: ENA24-balanced.csv

Analyze this dataset without downloading it

On Scigantic, IDLE-OO Camera Traps: Balanced Five-Set Species Classification Benchmark from LILA BC Camera Trap Images (Imageomics, BioCLIP 2) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Life Sciences datasets