KonIQ-10k Database

An ecologically valid image quality assessment database

KonIQ-10k is, at the time of publication, the largest IQA dataset to date consisting of 10,073 quality scored images. This is the first in-the-wild database aiming for ecological validity, with regard to the authenticity of distortions, the diversity of content, and quality-related indicators. Through the use of crowdsourcing, we obtained 1.2 million reliable quality ratings from 1,459 crowd workers, paving the way for more general IQA models.

We introduce a novel, deep learning model (KonCept512), to show an excellent generalization beyond the test set (0.921 SROCC), to the current state-of-the-art database LIVE-in-the-Wild (0.825 SROCC). The model derives its core performance from the InceptionResNet architecture, being trained at a higher resolution than previous models (512x384). A correlation analysis shows that KonCept512 performs similar to having 9 subjective scores for each test image.

Cite us

KonIQ-10k is freely available to the research community. If you use our database in your research, you can cite it as follows:

@article{koniq10k,
author={V. {Hosu} and H. {Lin} and T. {Sziranyi} and D. 
{Saupe}},
journal={IEEE Transactions on Image Processing},
title={KonIQ-10k: An Ecologically Valid Database for Deep 
Learning of Blind Image Quality Assessment},
year={2020},
volume={29},
pages={4041-4056}}

Alternative download

Downloads

Browse all images (online gallery)

5 Gb @ 1024x768 px

767 Mb @ 512x384 px

304 Kb download

389 Kb download

The scores file header

image_name: the image file name
c1-c5: number of ratings for each ACR value
c_total: total number of judgments per image
MOS: Mean Opinion Scores of the 5-point ACR
SD: Standard Deviation of the MOS
MOS_zscore: ACR scores are first normalised for each user by z-scoring the user ratings

Users exhibiting unusual scoring behaviour are removed. See our paper for more details on how we do reliability screening.