Share Email Print

Proceedings Paper

Adapting the ISO 20462 softcopy ruler method for online image quality studies
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

In this paper we address the problem of Image Quality Assessment of no reference metrics, focusing on JPEG corrupted images. In general no reference metrics are not able to measure with the same performance the distortions within their possible range and with respect to different image contents. The crosstalk between content and distortion signals influences the human perception. We here propose two strategies to improve the correlation between subjective and objective quality data. The first strategy is based on grouping the images according to their spatial complexity. The second one is based on a frequency analysis. Both the strategies are tested on two databases available in the literature. The results show an improvement in the correlations between no reference metrics and psycho-visual data, evaluated in terms of the Pearson Correlation Coefficient.

Paper Details

Date Published: 4 February 2013
PDF: 10 pages
Proc. SPIE 8653, Image Quality and System Performance X, 86530E (4 February 2013); doi: 10.1117/12.2007480
Show Author Affiliations
Peter D. Burns, Burns Digital Imaging (United States)
Jonathan B. Phillips, NVIDIA Corp. (United States)
Don Williams, Imaging Science Associates (United States)

Published in SPIE Proceedings Vol. 8653:
Image Quality and System Performance X
Peter D. Burns; Sophie Triantaphillidou, Editor(s)

© SPIE. Terms of Use
Back to Top