Div2k

Div2k专注于图像超分辨率任务，它包含RGB图像，并提供多种下采样比例（2、3、4、8倍）的训练集（800张）和验证集（100张）。数据来源于互联网，经过筛选保证图像质量和内容多样性，适用于图像超分辨率模型的训练和评估。该数据集仅供学术研究使用，并提供标准化数据操作和统一模型接口。

Div2K is dedicated to the image super-resolution task. It contains RGB images, and provides a training set (800 images) and a validation set (100 images) with multiple downsampling scales (2×, 3×, 4×, 8×). The dataset is sourced from the Internet, and has been filtered to ensure both image quality and content diversity, making it suitable for training and evaluating image super-resolution models. This dataset is solely for academic research use, and provides standardized data operations and unified model interfaces.

搜集汇总

背景与挑战

背景概述

Div2k是一个专注于图像超分辨率任务的RGB图像数据集，包含2K分辨率的高质量图像，提供2、3、4、8倍下采样比例的训练集（800张）和验证集（100张）。数据来源于互联网，经过精心筛选以确保图像质量和内容多样性，适用于图像超分辨率模型的训练和评估，且仅限于学术研究使用。