This is the training dataset for our AAAI2024 paper Iterative Token Evaluation and Refinement for Real-World Super-Resolution.You may use this dataset with our released codes in Github.
The datasets contains sub-image pairs of (LQ, HQ) patches.
HQ_subimages: high quality image patches extracted from DIV2K, DIV8K, Flicker2K, OutdoorSceneTraining and FFHQ.
LQ_subimages_X2: low quality patches with x2 downsampling
LQ_subimages_X4: low quality patches with x4… See the full description on the dataset page:
https://huggingface.co/datasets/chaofengc/ITER.
