Delft Image Quality Lab
 

IQ Lab
line

Introduction
People

 

Resources
line

A&A_dataset: The Aesthetics and visual Attention image dataset
EMMA: Database for emotion and mood recognition
i_QoE: A database for individual QoE analysis
Perceived Ringing
Eye-Tracking Release 1
Eye-Tracking Release 2
Interactions

Interactions

The eye-tracking data are publicly available to the research community. Please cite the following references if you use this database in your research.

  • J. Redi and I. Heynderickx, "TUD Image Quality Database: Interactions", http://mmi.tudelft.nl/iqlab/interactions.html.

  • J. Redi, H. Liu, R. Zunino, I. Heynderickx, "Interactions of visual attention and quality perception", IS&T/SPIE Electronic Imaging 2011 and Human Vision and Electronic Imaging XVI. 7865, 2011.

Download
Database: Interactions

The files are password protected. To get the password you can contact Judith Redi (j.a.redi@tudelft.nl)

Database Description

The interactions database was designed to invesigate on the deviations of quality scoring saliency from free looking saliency. In particular, the effect of two factor is of interest: the quality level of the images evaluated, and the distortion affecting them. Eye movements were recorded by means of a SensoMotoric Instruments GmbH Eye Tracker while 14 subjects were scoring the quality of 54 distorted stimuli with the Single Stimulus method with continuous numerical scaling. The stimuli used in the experiment consisted of several distorted versions of 6 original images selected from the LIVE database. These were selected such that we had a fair representation of different content, including images with and without a clear region of interest, images with and without humans in the picture, and images with various amounts of textured components. To evaluate the effect of distortion type and quality level, various distorted version of these original images were chosen from the LIVE database. Three kinds of distortions were considered, namely JPEG compression, White noise and Gaussian Blur. For each original image and distortion type, three different quality levels were selected: a highly distorted one, one with a medium distortion level, and one for which the applied distortion just slightly compromised the quality. The interaction database includes, together with the original stimuli, the corresponding fixation and saliency maps and the MOS obtained from the scoring task. The experimental procedure is described in more detail in the paper "Interactions of visual attention and quality perception".

 

This research group is part of the Interactive Intelligence group based in the Technical University of Delft
For questions or comments regarding this page, please contact Hani Alers