

The (VI (S) TA) Image Quality Toolbox ( VistaQualityTools ) is a Matlab Toolbox for full reference color (and also achromatic) image quality assessment based on divisive normalization models in DCT and wavelet domains. The general idea to assess the perceptual distance between two images is to compute the qnorm Euclidean distance in the image representation at the V1 visual cortex, as suggested in [Teo&Heeger, IEEE ICIP 1994]. This markedly differs from the Mean Square Error (2norm Euclidean measure in the spatial domain), as shown in [Pons99, Epifanio03] NEW! These ideas have been implemented in the wavelet domain in the new code associated to the paper [Laparra10a], to be included in a future version of VistaQualityTools. Results using the wavelet based measure outperform SSIM and VIF and are intuitively interpretable in a linear way. The updated version of the wavelet image quality measure is not included in the general toolbox, but can be downloaded from HERE. Citation: We have decided to make the library available to the research community free of charge. If you use VistaQualityTools in your research, we kindly ask that you reference this website: J. Malo, J. Gutiérrez, J. Muñoz and M. Simón. "VistaQualityTools: an image quality assessment toolbox for Matlab", http://www.uv.es/vista/vistavalencia/software/software.html
and the paper(s) associated to each algorithm.
Note that the package also contains some previously released public domain wavelet software authored by Eero P. Simoncelli, belonging to his MatlabPyrTools toolbox (http://www.cns.nyu.edu/~lcv/software.php). When using waveletbased functions in VistaQualityTools you shall aknowledge the author of MatlabPyrTools as well!.
Installation:
Basic features:
2norm Euclidean measure in a DCTbased divisive normalized
This demo applies different degradations to some particular image leading to distorted images with the same MSE but quite
Some Results: The suitability of a distortion measure can be seen by looking at the correlation between subjective distortions (e.g. measured by Mean Opinion Score MOS) and the values predicted by the distortion measure. In the examples below we compare this agreement for different distortion measures. RMSE and our DCTbased measure for images corrupted by white noise 
RMSE 
MPE DCTbased measure 
RMSE, Structural Similarity Index SSIM (Wang et al. IEEE Tr.Im.Proc. 2004) and our Waveletbased measure for images degraded by JPEG, JPEG2000 and fast fading distortion 
RMSE 
SSIM 
MPE WaveletBased Measure 
Download VistaQualityTools ! 

