Image quantization and digital halftoning are fundamental problems in compu
ter graphics, which arise when displaying high-color images on non-truecolo
r devices. Both steps am generally performed sequentially and, in most case
s, independent of each other Color quantization with a pixel-wise defined d
istortion measure and the dithering process with its local neighborhood opt
imize different quality criteria or, frequently, follow a heuristic without
reference to any quality measure.
In this paper we propose a new method to simultaneously quantize and dither
color images. The method is based on a vigorous cost-function approach whi
ch optimizes a quality criterion, derived from a generic model of human per
ception. A highly efficient algorithm for optimization based on a multiscal
e method is developed for the dithered color quantization cost function. Th
e quality criterion and the optimization algorithms are evaluated on a repr
esentative set of artificial and real-world images us well as on a collecti
on of icons. A significant image quality improvement is observed compared t
o standard color reduction approaches.