Abstract

Colors sampled in an array of pixels were converted into various luminance–chrominance representations, and the chrominance values were spatially subsampled to achieve a compression of the data. For viewing, the chrominance values were reinterpolated and then transformed back to RGB. For each of a range of image types and color spaces, we varied the chromatic sample spacing to determine how much compression would be possible before perceptual artifacts appeared, as determined by a panel of individual viewers. The spacing at which an image was mistaken half the time for a remembered original depended on image content and also on the color representation. Some color spaces, such as hue–lightness–saturation, a polar-coordinate color space, are inherently ill suited for chrominance averaging and produce noticeable artifacts on almost any image. Likewise, some images cannot be subsampled satisfactorily, regardless of the color space used. In general, however, images with 16-fold reduction in the amount of chrominance information (that is, a factor of 4 in each spatial direction) are still satisfactory in appearance.

© 1990 Optical Society of America

Visual interactions with luminance and chromatic stimuli

G. R. Cole, C. F. Stromeyer, and R. E. Kronauer
J. Opt. Soc. Am. A 7(1) 128-140 (1990)

Critical color differences determined with a visual search task

Allen L. Nagy and Robert R. Sanchez
J. Opt. Soc. Am. A 7(7) 1209-1217 (1990)

Perceptually uniform color space for image signals including high dynamic range and wide gamut

Muhammad Safdar, Guihua Cui, Youn Jin Kim, and Ming Ronnier Luo
Opt. Express 25(13) 15131-15151 (2017)

Visual search for color differences with foveal and peripheral vision

Allen L. Nagy, Robert R. Sanchez, and Thomas C. Hughes
J. Opt. Soc. Am. A 7(10) 1995-2001 (1990)

Orientation and spatial-frequency discrimination for luminance and chromatic gratings

Michael A. Webster, Karen K. De Valois, and Eugene Switkes
J. Opt. Soc. Am. A 7(6) 1034-1049 (1990)