{"url":"/dataset/div2k","name":"DIV2K","full_name":null,"description_markdown":"**DIV2K** is a popular single-image super-resolution dataset which contains 1,000 images with different scenes and is splitted to 800 for training, 100 for validation and 100 for testing. It was collected for NTIRE2017 and NTIRE2018 Super-Resolution Challenges in order to encourage research on image super-resolution with more realistic degradation. This dataset contains low resolution images with different types of degradations. Apart from the standard bicubic downsampling, several types of degradations are considered in synthesizing low resolution images for different tracks of the challenges. Track 2 of NTIRE 2017 contains low resolution images with unknown x4 downscaling. Track 2 and track 4 of NTIRE 2018 correspond to realistic mild ×4 and realistic wild ×4 adverse conditions, respectively. Low-resolution images under realistic mild x4 setting suffer from motion blur, Poisson noise and pixel shifting. Degradations under realistic wild x4 setting are further extended to be of different levels from image to image.\r\n\r\nSource: [Unsupervised Image Super-Resolution with an Indirect Supervised Path](https://arxiv.org/abs/1910.02593)","description_withheld":null,"homepage":"https://data.vision.ee.ethz.ch/cvl/DIV2K/","introduced_date":"2017-01-01","introduced_date_note":null,"introduced_by":{"paper":null,"title":"NTIRE 2017 Challenge on Single Image Super-Resolution: Dataset and Study","first_author":null,"url":"https://doi.org/10.1109/CVPRW.2017.150"},"license":{"name":"Custom (research-only)","url":"https://data.vision.ee.ethz.ch/cvl/DIV2K/"},"modalities":[{"name":"Images","url":"/datasets/modality/images"}],"tasks":[{"name":"Denoising","url":"/task/denoising","datasets_with_task":"/datasets/task/denoising"},{"name":"Image Super-Resolution","url":"/task/image-super-resolution","datasets_with_task":"/datasets/task/image-super-resolution"},{"name":"JPEG Artifact Correction","url":"/task/jpeg-artifact-correction","datasets_with_task":"/datasets/task/jpeg-artifact-correction"},{"name":"Image Rescaling","url":"/task/image-rescaling","datasets_with_task":"/datasets/task/image-rescaling"},{"name":"Jpeg Compression Artifact Reduction","url":"/task/jpeg-compression-artifact-reduction","datasets_with_task":"/datasets/task/jpeg-compression-artifact-reduction"}],"languages":[{"name":"Chinese","url":"/datasets/language/chinese"}],"variants":["DIV2K","DIV2K val - 16x upscaling","DIV2K val - 2x upscaling","DIV2K val - 4x upscaling"],"data_loaders":[{"repo":"https://github.com/tensorflow/datasets","url":"https://www.tensorflow.org/datasets/catalog/div2k","frameworks":["tf","jax"]},{"repo":"https://github.com/eugenesiow/super-image-data","url":"https://github.com/eugenesiow/super-image-data","frameworks":["pytorch"]}],"num_papers_in_archive":654,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/image-super-resolution-on-div2k-val-4x","task":"Image Super-Resolution","dataset_variant":"DIV2K val - 4x upscaling","rows":21,"metrics":["LPIPS","PSNR","SSIM","DISTS","LRPSNR","NIQE"],"first_row_in_archive_order":{"model":"AESOP","paper":"/paper/auto-encoded-supervision-for-perceptual-image","metrics":{"DISTS":"0.0459","LPIPS":"0.0893","PSNR":"29.137","SSIM":"0.8023"},"code_links":[{"title":"2minkyulee/aesop-auto-encoded-supervision-for-perceptual-image-super-resolution","url":"https://github.com/2minkyulee/aesop-auto-encoded-supervision-for-perceptual-image-super-resolution"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/denoising-on-div2k","task":"Denoising","dataset_variant":"DIV2K","rows":1,"metrics":["Average PSNR (dB)"],"first_row_in_archive_order":{"model":"DRUnet_Poisson_0.01","paper":"/paper/generalized-recorrupted-to-recorrupted-self","metrics":{"Average PSNR (dB)":"33.92"},"code_links":[{"title":"deepinv/deepinv","url":"https://github.com/deepinv/deepinv"},{"title":"bemc22/GeneralizedR2R","url":"https://github.com/bemc22/GeneralizedR2R"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/image-super-resolution-on-div2k-val-16x","task":"Image Super-Resolution","dataset_variant":"DIV2K val - 16x upscaling","rows":1,"metrics":["PSNR","SSIM"],"first_row_in_archive_order":{"model":"ABPN","paper":"/paper/image-super-resolution-via-attention-based","metrics":{"PSNR":"24.38","SSIM":"0.641"},"code_links":[{"title":"Holmes-Alan/ABPN","url":"https://github.com/Holmes-Alan/ABPN"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/generalized-recorrupted-to-recorrupted-self","title":"Generalized Recorrupted-to-Recorrupted: Self-Supervised Learning Beyond Gaussian Noise","date":"2024-12-05","rows_on_this_dataset":1,"code_links":2,"syntology":null},{"paper":"/paper/auto-encoded-supervision-for-perceptual-image","title":"Auto-Encoded Supervision for Perceptual Image Super-Resolution","date":"2024-11-28","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":6,"samples_ran":2,"samples_unverified":4,"pointer_only_for_licence":6,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/residual-conditioned-optimal-transport","title":"Residual-Conditioned Optimal Transport: Towards Structure-Preserving Unpaired and Paired Image Restoration","date":"2024-05-05","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":12,"samples_ran":11,"samples_unverified":1,"pointer_only_for_licence":11,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/boosting-flow-based-generative-super","title":"Boosting Flow-based Generative Super-Resolution Models via Learned Prior","date":"2024-03-16","rows_on_this_dataset":2,"code_links":1,"syntology":null},{"paper":"/paper/image-restoration-through-generalized","title":"Image Restoration Through Generalized Ornstein-Uhlenbeck Bridge","date":"2023-12-16","rows_on_this_dataset":1,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":2,"samples_ran":0,"samples_unverified":2,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/implicit-diffusion-models-for-continuous","title":"Implicit Diffusion Models for Continuous Super-Resolution","date":"2023-03-29","rows_on_this_dataset":7,"code_links":1,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":10,"samples_ran":5,"samples_unverified":5,"pointer_only_for_licence":10,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/local-implicit-normalizing-flow-for-arbitrary","title":"Local Implicit Normalizing Flow for Arbitrary-Scale Image Super-Resolution","date":"2023-03-09","rows_on_this_dataset":2,"code_links":3,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":5,"samples_ran":2,"samples_unverified":3,"pointer_only_for_licence":5,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/perception-oriented-single-image-super","title":"Perception-Oriented Single Image Super-Resolution using Optimal Objective Estimation","date":"2022-11-24","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/flexible-style-image-super-resolution-using","title":"Flexible Style Image Super-Resolution using Conditional Objective","date":"2022-01-13","rows_on_this_dataset":2,"code_links":1,"syntology":null},{"paper":"/paper/learning-continuous-image-representation-with","title":"Learning Continuous Image Representation with Local Implicit Image Function","date":"2020-12-16","rows_on_this_dataset":1,"code_links":2,"syntology":null},{"paper":"/paper/multi-step-reinforcement-learning-for-single","title":"Multi-Step Reinforcement Learning for Single Image Super-Resolution","date":"2020-07-28","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/srflow-learning-the-super-resolution-space","title":"SRFlow: Learning the Super-Resolution Space with Normalizing Flow","date":"2020-06-25","rows_on_this_dataset":1,"code_links":8,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":3,"samples_ran":2,"samples_unverified":1,"pointer_only_for_licence":3,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/image-super-resolution-via-attention-based","title":"Image Super-Resolution via Attention based Back Projection Networks","date":"2019-10-10","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/enhanced-deep-residual-networks-for-single","title":"Enhanced Deep Residual Networks for Single Image Super-Resolution","date":"2017-07-10","rows_on_this_dataset":1,"code_links":45,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":4,"samples_ran":4,"samples_unverified":0,"pointer_only_for_licence":2,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":7,"samples_harvested":42,"samples_ran":26,"samples_unverified":16,"pointer_only_for_licence":37,"papers_with_no_sample_that_ran":1,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}