KADID10k
KADID10k is a large-scale dataset designed for the evaluation of image quality assessment (IQA) algorithms. It was introduced to address the limitations of existing datasets, which often lack diversity in terms of content, distortion types, and distortion levels. The dataset consists of 10,000 high-quality reference images, each paired with five distorted versions, resulting in a total of 60,000 images. The distortions include various types such as JPEG compression, Gaussian blur, white noise, and more, applied at different levels to simulate real-world image degradation scenarios. KADID10k is notable for its comprehensive coverage of distortion types and levels, making it a valuable resource for researchers and developers in the field of image processing and computer vision. The dataset has been used in numerous studies to benchmark the performance of IQA algorithms, contributing to the advancement of image quality assessment techniques.