Dataset Card for Caltech-UCSD Birds-200-2011 (CUB-200-2011)

Dataset Details

Dataset Description

The CUB-200-2011 dataset is designed for fine-grained bird species classification. It contains 11,788 images of 200 bird species, each labeled with its corresponding category. The dataset is widely used in fine-grained image classification and computer vision research.

Dataset Sources

• Homepage: https://www.vision.caltech.edu/datasets/cub_200_2011/
• Paper: Wah, C., Branson, S., Welinder, P., Perona, P., & Belongie, S. (2011). The caltech-ucsd birds-200-2011 dataset.

Dataset Structure

Each sample in the dataset contains:

• image: A variable-sized RGB image

• label: An integer between 0 and 199, representing the bird species

Total images: 11,788

Classes: 200 (each corresponding to a bird species)

Splits:

• Train: 5,994 images

• Test: 5,794 images

Image specs: Variable sizes, RGB

Example Usage

Below is a quick example of how to load this dataset via the Hugging Face Datasets library.

from datasets import load_dataset 

# Load the dataset 
dataset = load_dataset("randall-lab/cub200", split="train", trust_remote_code=True)
# dataset = load_dataset("randall-lab/cub200", split="test", trust_remote_code=True) 

# Access a sample from the dataset 
example = dataset[0] 
image = example["image"] 
label = example["label"] 

image.show() # Display the image 
print(f"Label: {label}")

Citation

BibTeX:

@article{wah2011caltech, title={The caltech-ucsd birds-200-2011 dataset}, author={Wah, Catherine and Branson, Steve and Welinder, Peter and Perona, Pietro and Belongie, Serge}, year={2011}, publisher={California Institute of Technology} }