A Principal Distance Conastraint Error Diffusion Algorithm for the Homogeneous Dot Distributions

초록

The perceived quality of the halftoned image strongly depends on the spatial distribution of the binary dots. Various error diffusion algorithms have been proposed for realizing the homogeneous dot distribution in the highlight and shadow regions. However, they are computationally expensive and/or require large memory space. This paper presents a new threshold modulated error diffusion algorithm for the homogeneous dot distribution. The proposed method is applied exactly same as the Floyd-Steinberg’s algorithm except the thresholding process. The threshold value is modulated based on the difference between the distance to the nearest minor pixel, “minor pixel distance”, and the principal distance. To do so, calculation of the minor pixel distance is needed for every pixel. But, it is quite time consuming and requires large memory resources. In order to alleviate this problem, “the minor pixel offset array” that transforms the 2-dimensional history of minor pixels into the 1-dimensional codes is proposed. The proposed algorithm drastically reduces the computational load and memory spaces needed for calculation of the minor pixel distance.