ImageInWords: Unlocking Hyper-Detailed Image Descriptions

arXiv: https://arxiv.org/abs/2405.02793

Please visit the webpage for all the information about the IIW project, data, visualizations, and much more. The data can be downloaded directly from the datasets/ folder, as well as from Huggingface (see below).

Please reach out to [email protected] for thoughts/feedback/questions/collaborations.

License: CC-BY-4.0

Other resources

🤗Hugging Face🤗

IIW-Benchmark Eval Dataset

from datasets import load_dataset

# `name` can be one of: IIW-400, DCI_Test, DOCCI_Test, CM_3600, LocNar_Eval
# refer: https://github.com/google/imageinwords/blob/main/datasets/README.md
dataset = load_dataset('google/imageinwords', token=None, name="IIW-400", trust_remote_code=True)

Dataset-Explorer

Cite

If you use our data or refer to our work, please include the following citation

@misc{garg2024imageinwords,
      title={ImageInWords: Unlocking Hyper-Detailed Image Descriptions}, 
      author={Roopal Garg and Andrea Burns and Burcu Karagol Ayan and Yonatan Bitton and Ceslee Montgomery and Yasumasa Onoe and Andrew Bunner and Ranjay Krishna and Jason Baldridge and Radu Soricut},
      year={2024},
      eprint={2405.02793},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}