Real-BlindSR Dataset

Realistic benchmark datasets are crucial for providing a consistent measurement baseline for comparing different BlindSR methods and testing their generalization ability in real-world scenarios. However, the evaluation of BlindSR methods is still limited by the lack of a common dataset that conforms to realistic degradation scenarios. We construct a benchmark dataset that follows real scenarios, which reflects the real-world BlindSR problem more accurately than existing synthetic datasets. The Real-BlindSR dataset includes diverse degradation types frequently observed in realistic situations, such as different downsampling scales. This dataset offers a comprehensive and realistic benchmark for assessing the performance of BlindSR methods.The Real-BlindSR Dataset we built is mainly used to evaluate the BlindSR task, which includes LR-HR paired images of three different scales of X2, X4, and X8, and includes compression from different Internet platforms and processing of real-world degradation three different tracks.