Share Email Print
cover

Proceedings Paper

Economical subject-selective metrics of perceptible color image quality
Author(s): Vitali V. Gavrik
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Economical quality metrics for electronic imaging are to obey natural psychophysical laws. The perceptible image quality has proved to be related subconsciously to few dominant objects that could be standardized. The brain response is a product of their single-stimulus quality functions whose mathematics is typical to other psychophysical response-stimulus dependencies. They contain a high-quality tolerance plateau and one or two linear portions of quality decrease with the log-stimulus variation. Every stimulus is an optical property of an image detail, simply measurable and directly perceived by vision. Transfer properties of imaging system are to provide the stimuli to meet their psychophysical tolerances. In spite of much smaller optical differences, perceptible under some artificial non-imaging conditions (e.g., colorimetry), most of scene objects have long high-quality plateau. Few dominant image details make the large photometric distortions in all the rest not attract observers' attention. Intentional drawing the attention to some image features may involve other details into evaluation. The stimulus quality functions can be either extracted by an iteration procedure from the multi-stimulus data on a representative set of images or obtained in a single-stimulus experiment. Instead of the numerous rough correlations to arbitrary chosen integral features of different images, the exhaustive description can directly and accurately predict the perceptible color image quality, issuing from a few simple optical measurements, and appears to be generalized on either quantifiable produce features.

Paper Details

Date Published: 7 September 1998
PDF: 12 pages
Proc. SPIE 3409, Electronic Imaging: Processing, Printing, and Publishing in Color, (7 September 1998); doi: 10.1117/12.324096
Show Author Affiliations
Vitali V. Gavrik, Consultant (Germany)


Published in SPIE Proceedings Vol. 3409:
Electronic Imaging: Processing, Printing, and Publishing in Color
Jan Bares, Editor(s)

© SPIE. Terms of Use
Back to Top