okhater/NTIRE_LLE_2025

---

license: apache-2.0

NTIRE 2025 Low Light Image Enhancement Challenge Dataset

Overview

The NTIRE 2025 dataset is crafted to benchmark low-light image enhancement algorithms, featuring a diverse set of challenging low-light conditions. It is structured to support both the development and evaluation phases of the challenge.

Dataset Composition

• Training Set: 219 images with paired low-light inputs and corresponding ground truth images.
• Validation Set: 46 images provided as low-light inputs.
• Test Set: 30 images provided as low-light inputs.

Folder Structure

The dataset is organized as follows:

NTIRE2025/
  ├── train/ │
    └── samples/ # Paired samples for training
      │ ├── sample_001/ │
        │ ├── GT.jpg # Ground Truth image │
        │ └── Input.jpg # Low Light image │
      ├── sample_002/ │
        │ ├── GT.jpg │
        │ └── Input.jpg
      │ └── ...
  ├── val/
    │ └── samples/ # Samples for validation (low-light images only)
      │ ├── sample_001/
        │ │ └── Input.jpg
        │ └── ...
  └── test/
    └── samples/ # Samples for testing (low-light images only)
      ├── sample_001/
      │ └── Input.jpg
        └── ...

Data Source

The dataset is available via designated download links provided by the organizers. Participants can access the dataset using the provided Google Drive links or via Baidu WangPan (password: 2025). For more details on the challenge, please visit the NTIRE 2025 Challenge Page.

Code Samples for Data Loading

• For data loaders and preprocessing examples, see "load_data_loaders.py".
• For dataset loading and inspection examples, see "load_dataset.py".

Citation

If you use this dataset in your research, please cite the following paper:

@article{codalab_competitions_JMLR,
  author  = {Adrien Pavao and Isabelle Guyon and Anne-Catherine Letournel and Dinh-Tuan Tran and Xavier Baro and Hugo Jair Escalante and Sergio Escalera and Tyler Thomas and Zhen Xu},
  title   = {CodaLab Competitions: An Open Source Platform to Organize Scientific Challenges},
  journal = {Journal of Machine Learning Research},
  year    = {2023},
  volume  = {24},
  number  = {198},
  pages   = {1--6},
  url     = {http://jmlr.org/papers/v24/21-1436.html}
}

Additionally, if you find the dataset card useful in your workflow, please consider citing the dataset card as follows:

@misc{NTIRE2025_dataset_card,
  author       = {Omar Khater},
  title        = {NTIRE 2025 Low Light Image Enhancement Challenge Dataset Card},
  howpublished = {\url{https://huggingface.co/datasets/okhater/NTIRE_LLE_2025}},
  year         = {2025},
  note         = {Dataset card curated for clarity and efficient access to the NTIRE 2025 challenge data}
}