The purpose of this page is to make available one subjective quality assessment from IRCCyN/IVC. The experiment was conducted in our lab in normalized conditions.

Content

5 original grayscale images, 210 distorted images were generated from noise addition in 6 different frequency (Fourier) sub-bands of various frequency range, with 7 embedding strengths.

Details on the spreadsheet.

Marked images are entitled "Marked_boats_X_Y.bmp", where 'X' stands for the embedding sub band index in a Perceptual Channel Decomposition (PCD), splitting the Fourier frequency stectrum into 17 overlapping sub-bands, see the paper referenced at the bottom of the page for further details on the perceptual decomposition (the PCD is depicted on the figure below where every sub-band is indexed) and 'Y' is the embedding strength. A Gaussian noise was independently modulated within one PCD sub band, and spread over the whole sub-band.

Only subbands number 2, 3, 6, 7, 12 and 13 (see index on Figure 1) were used during this experiment.

PCD

Figure 1. Perceptual Channel Decomposition with indexed sub-bands.

The purpose of this experiment was to highlight the incoherencies of objective quality metrics when distortions occur in a different frequency range. Subjective evaluations were made at viewing distance of 6 times the screen height using a DSIS (Double Stimulus Impairment Scale) method with 5 categories and 7 observers. Distortions for each processing and each image have been optimised in order to uniformly cover the subjective scale.

Download

Download the IVC-FourierSB database (52 MB zip file)
A Microsoft Excel spreadsheet, giving all subjective quality scores is included in the above archive.

Dissemination

This database was used in

M. Carosi, V. Pankajakshan, F. Autrusseau, "Toward a simplified perceptual quality metric for watermarking applications", Proceedings of the SPIE Electronic Imaging, conf. 7542, January 2010, San Jose, CA, USA.

Reference

F. Autrusseau, P. Le Callet, "A robust image watermarking technique based on quantization noise visibility thresholds",Elsevier Signal Processing, Volume 87, Issue 6, June 2007, Pages 1363-1383.