KonIQ-10k stands for Konstanz Image Quality 10k Database, which is a large IQA dataset consisting of 10,073 high-quality rated images. It is the first field database for ecological validity, involving distortion authenticity, content diversity, and quality-related indicators.

This dataset obtains 1.2 million reliable quality ratings from 1,459 crowdsourced workers through crowdsourcing, facilitating the study of more general IQA models.