Multi-Class Anisotropic Electrostatic Halftoning

• Published in: Computer graphics forum 2012-09, Vol.31 (6), p.1924-1935
• Main Authors: Schmaltz, C., Gwosdek, P., Weickert, J.
• Format: Article
• Language: English
• Subjects: Algorithms; Analysis; Anisotropy; Computer graphics; Electrostatics; halftoning; hatching; I.3.3 [Computer Graphics]: Picture/Image Generation-Display Algorithms; I.3.5 [Computer Graphics]: Computational Geometry and Object Modelling-Physically Based Modelling; I.3.8 [Computer Graphics]: Applications-Dithering and Stippling; Image processing systems; multi-class; multi-size; second-order screening; stippling; Studies
• ISSN: 0167-7055; 1467-8659
• DOI: 10.1111/j.1467-8659.2012.03072.x

Electrostatic halftoning, a sampling algorithm based on electrostatic principles, is among the leading methods for stippling, dithering and sampling. However, this approach is only applicable for a single class of dots with a uniform size and colour. In our work, we complement these ideas by advanced features for real‐world applications. We propose a versatile framework for colour halftoning, hatching and multi‐class importance sampling with individual weights. Our novel approach is the first method that globally optimizes the distribution of different objects in varying sizes relative to multiple given density functions. The quality, versatility and adaptability of our approach is demonstrated in various experiments. Electrostatic halftoning, a sampling algorithm based on electrostatic principles, is among the leading methods for stippling, dithering and sampling. However, this approach is only applicable for a single class of dots with a uniform size and colour. In our work, we complement these ideas by advanced features for real‐world applications. We propose a versatile framework for colour halftoning, hatching and multi‐class importance sampling with individual weights. Our novel approach is the first method that globally optimizes the distribution of different objects in varying sizes relative to multiple given density functions. The quality, versatility and adaptability of our approach is demonstrated in various experiments.