Estimation of Gray level Dependent Error Diffusion Kernels Based on Genetic Algorithm

초록

Error diffusion kernel that determines amount of error propagation to neighboring pixels is one of major factors affecting image quality of halftoned image. Recently, utilization of multiple kernels has been proposed to improve homogeneity of dot distribution and reduce unpleasing artifacts. This paper proposes a new technique for estimating gray level dependent error diffusion kernels. In the proposed technique, kernels are estimated by optimizing a correlation coefficient defined for radial power spectrum and blue noise profile. Genetic algorithm is chosen as an optimization tool.