Color halftoning using binary database

Abstract

A new halftoning method for color and monochrome images is proposed to achieve optimal image quality as well as fast processing time. The halftoning method is very simple in computation, which indexes the image pixel values to a binary database as proposed by Sullivan et al. [1]. The binary database is a set of dot profiles which are sequentially generated from a seed pattern which is well-distributed by circular convolution with the contrast sensitivity function (CSF), in order that the dot profiles may have a visual-optimized property as well as a correlated property with each other. The circular convolution is useful technique to reduce the memory size for the database. The proposed color halftoning method uses the same database as monochrome halftoning, but each color component has different indexing scheme with a circularly shifted indexing to remove the correlation between color components. The simple structure and small memory requirement allow the fastest processing time and the simplest hardware solution. The halftone image quality is better than that of ordered dither [2], and if preprocessing of an edge enhancement is additionally applied, then the image quality is similar to that of error diffusion [3].