RGB, a nonuniform color space, is almost universally accepted by the image
processing community as the means for representing color. On the other hand
, perceptually uniform spaces, such as L*a"b*, as well as approximately-uni
form color spaces, such as HSV, exist, in which measured color differences
are proportional to the human perception of such differences.
This paper compares RGB with L*a*b* and HSV in terms of their effectiveness
in color texture analysis, There has been a limited but increasing amount
of work on the color aspects of textured images recently. The results have
shown that incorporating color into a texture analysis and recognition sche
me can be very important and beneficial.
The presented methodology uses a family of Gabor filters specially tuned to
measure specific orientations and sizes within each color texture. Effecti
veness is measured by classification performance of each color space, as we
ll as by classifier-independent measures. Experimental results are obtained
with a variety of color texture images. Percuptually uniform spaces are sh
own to outperform RGB in many cases.