A novel version of the popular SSIM image quality assessment technique, specifically designed to yield uniform MOS prediction in presence of different types of image distortions is presented. Response equalization is obtained through the polynomial combination of a basic SSIM metric with an auxiliary metric characterized by a diverse sensitivity. Accuracy and consistency of this composite index, called E-SSIM, are demonstrated with experiments conducted on two independent archives annotated with MOS values. © 2011 University of Zagreb.
Equalized structural similarity index for image quality assessment / L., Capodiferro; F., Mangiatordi; DI CLAUDIO, Elio; Iacovitti, Giovanni. - STAMPA. - (2011), pp. 420-424. (Intervento presentato al convegno 7th International Symposium on Image and Signal Processing and Analysis, ISPA 2011 tenutosi a Dubrovnik; Croatia nel 4 September 2011 through 6 September 2011).
Equalized structural similarity index for image quality assessment
DI CLAUDIO, Elio;IACOVITTI, Giovanni
2011
Abstract
A novel version of the popular SSIM image quality assessment technique, specifically designed to yield uniform MOS prediction in presence of different types of image distortions is presented. Response equalization is obtained through the polynomial combination of a basic SSIM metric with an auxiliary metric characterized by a diverse sensitivity. Accuracy and consistency of this composite index, called E-SSIM, are demonstrated with experiments conducted on two independent archives annotated with MOS values. © 2011 University of Zagreb.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
