Unlike the error diffusion method, the dot diffusion method for digital hal
ftoning has the advantage of pixel-level parallelism, However, image qualit
y offered by error diffusion is still regarded as superior to most of the o
ther known methods. In this paper, we show how the dot diffusion method can
be improved by optimization of the so-called class matrix, By taking the h
uman visual characteristics into account we show that such optimization con
sistently results in images comparable to error diffusion, without sacrific
ing the pixel-level parallelism. Adaptive dot diffusion is also introduced
and then a mathematical description of dot diffusion is derived. Furthermor
e, inverse halftoning of dot diffused images is discussed and two methods a
re proposed. The first one uses projection onto convex sets (POCS) and the
second one uses wavelets. Of these methods, the wavelet method does not mak
e use of the knowledge of the class matrix, Embedded multiresolution dot di
ffusion is also discussed, which is useful for rendering at different resol
utions and transmitting images progressively.