Affiliation:
1. CGNIP, Gatineau, Canada
2. School of Computer Science, Carleton University, Ottawa, Canada
Abstract
This article presents a new and efficient automatic method for structure-preserving stippling. The core idea is to concentrate on structure preservation by using a priority-based scheme that treats extremal pixels first and preferentially assigns positive error to lighter pixels and negative error to darker pixels, emphasizing contrast. The use of a nonlinear spatial function to shrink or exaggerate errors implicitly provides global adjustment of density. Personal adjustment respects contrast and hence allows people to preserve structure even with few stipples. Beyond the advantage of good structure preservation, the algorithm provides many variations to extend personal stippling to other artistic styles. In addition, it is demonstrated that variations on priority-based schemes, by a multiple-stage process, can provide flexibility to promote different kinds of interesting features. This article explores a variety of stylized effects, including heightening, scratchboard, and line drawing, all within the unifying framework of stippling.