Visual Quality Assessment Databases, MMSP Konstanz
  • Home
  • VQA Databases
    • KonViD-150k Database
    • KonViD-1k Database
    • KonIQ-10k Database
    • KonIQ++ Database
    • KADID-10k Database
    • IQA-Experts-300
    • KonPatch-30k Database
    • KoSMo-1k Database
    • StudyMB 2.0 Database
    • Picturewise JND Data
    • KonJND-1k database
    • KonFiG-IQA Database
    • GFIQA-20k Database
  • About
  • Contact

KonIQ-10k IQA 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}}
Picture
Picture
Picture
Alternative download

Downloads

Browse all images (online gallery)
images Full (.zip)
5 Gb @ 1024x768 px
images small (.zip)
767 Mb @ 512x384 px
SCORES (.zip)
304 Kb download
indicators (.zip)
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.

Example images, with indicators

Mean
Opinion​
Score
Picture
MOS: 1.13
Picture
MOS: 1.97
Picture
MOS: 2.71
Picture
MOS: 3.52
Picture
MOS: 4.31
Brightness
Picture
Brightness: 0.03
Picture
Brightness: 0.20
Picture
Brightness: 0.41
Picture
Brightness: 0.60
Picture
Brightness: 0.80
Colorfulness
Picture
Colorfulness: 0
Picture
Colorfulness: 0.13
Picture
Colorfulness: 0.24
Picture
Colorfulness: 0.36
Picture
Colorfulness: 0.48
Contrast
Picture
Contrast: 0.06
Picture
Contrast: 0.14
Picture
Contrast: 0.23
Picture
Contrast: 0.31
Picture
Contrast: 0.40
Sharpness
Picture
Sharpness: 5.58
Picture
Sharpness: 10.19
Picture
Sharpness: 15.27
Picture
Sharpness: 20.54
Picture
Sharpness: 25.63
Imprint | Privacy policy | About | Contact
Copyright "VQA Group at Universität Konstanz" © 2019-2022
  • Home
  • VQA Databases
    • KonViD-150k Database
    • KonViD-1k Database
    • KonIQ-10k Database
    • KonIQ++ Database
    • KADID-10k Database
    • IQA-Experts-300
    • KonPatch-30k Database
    • KoSMo-1k Database
    • StudyMB 2.0 Database
    • Picturewise JND Data
    • KonJND-1k database
    • KonFiG-IQA Database
    • GFIQA-20k Database
  • About
  • Contact