Large-scale Ideal Ultra high definition 4K Version 2 (LIU4K-v2) Dataset

                        

Dataset Superiorities

LIU4K-v2 has several unprecedented superiorities as follows,

  • High-resolution definition.
    • Compared to previous datasets, the resolutions of the images in our dataset are at least 3K, most of which belong to 4K-6K, larger than those in previous datasets, which offers abundant materials for testing and evaluating the performance in 4K/8K display devices.
  • Large-scale.
    • Our dataset is large-scale. Our dataset includes 1600 high-resolution training images and 400 high-resolution validation images, which are much more than those in previous datasets. Thus, training and evaluation based on LIU4K are more comprehensive and balanced.
  • Diversified and complex contents.
    • Our dataset includes very diversified kinds of backgrounds and objects and also is diverse and complex in low-level signal distribution.
  • High visual quality.
    • Due to its high high-resolution definition and the diversified and complex contents, images in our LIU4K dataset has high visual qualtiy.

Dataset Overview

Figure. 1. Example training set images sampled from LIU4K-v2.

Download

  • Datasets
    Google Drive: Train, Validation
    Baiduyun: Train (extracted code: avzb), Validation (extracted code: bta6)
    PKU Drive: Train, Validation
    All the images are collected fom the Internet, and these images are published under the CC0 License (https://creativecommons.org/share-your-work/public-domain/cc0/). This means basically you can use these images for any purpose. Since the images are collected by us, we ask you to kindly cite the following paper if you are using this dataset.

    • Citation

    @ARTICLE{Liu4K, author={J. {Liu} and D. {Liu} and W. {Yang} and S. {Xia} and X. {Zhang} and Y. {Dai}}, journal={IEEE Transactions on Image Processing}, title={A Comprehensive Benchmark for Single Image Compression Artifact Reduction}, year={2020}, volume={29}, number={}, pages={7845-7860}, }