Subjective quality assessment - "Perceptually Optimized Watermarking" database

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

The "Perceptually Optimized Watermarking" (POW) database aims at tracking the visibility threshold of an image watermarking method.
An alternative title for this database would be: "2AFC evaluation of Fourier watermarked images/documents".

Content

9 original color images were used, 243 distorted images were generated with 3 different embedding algorithms, for each image (and each algorithm), 9 embedding strengths were used (9x9x3=243).

Details on this database can be found in the citation below. Subjective evaluations were made at viewing distance of 6 times the screen height using a 2AFC (Two Alternative Forced Choice) protocol. Distortions for each processing and each image have been optimised in order to uniformly cover the visibility scale.

37 observers were enrolled in this experiment. An outlier detection method was performed, the dendrogram is plotted here. Overall, 6 observers were discarded.


Download

Download the IVC-POW database (108 MB 7z file)
A Microsoft Excel spreadsheet, giving all subjective quality scores is included in the above archive.

How to cite the IVC-POW database

If you use BibTeX, here's the citation:

	@article{TIFS2014,
	    author = {M. Urvoy and D. Goudia and F. Autrusseau},
	    title = {Perceptual DFT watermarking with improved detection and robustness to geometrical distortions},
	    journal = {in IEEE Transactions on Information Forensics and Security, IEEE-TIFS},
	    year = {2014},
	    volume = {9(7)},
	    pages = {1108-1119},
	    note = {ISSN: 1556-6013},
	    url = {http://www.irccyn.ec-nantes.fr/~autrusse/Papers/TIFS2014.pdf},
	    doi = {http://dx.doi.org/10.1109/TIFS.2014.2322497}
	}

If you don't use BibTeX, use simply:

	M. Urvoy, D. Goudia, F. Autrusseau, 
		"Perceptual DFT watermarking with improved detection and robustness to geometrical distortions", 
		in IEEE Transactions on Information Forensics and Security, IEEE-TIFS, 9(7), pp. 1108-1119, ISSN 1556-6013, Jul. 2014

This work was supported by the 'StreamMaster' OSEO research project.