Image Data Field-Based Method for Stippling

Tao Wu

doi:10.1109/ACCESS.2019.2909159

Return

Image Data Field-Based Method for Stippling

2019-01-01

Tao Wu *

Junjie Yang

Xue Qin

DOI：10.1109/ACCESS.2019.2909159

Original

Save

View PDF

Abstract

En 中文

Computer-based generative art is in the ascendant and widely applied in media contexts such as web design, game resource, and digital entertainment. The process of creating non-photorealistic rendering images can be enjoyable if some useful and efficient methods are involved. This paper presents an automatic and fast method for digital stippling, which produces stipple renderings from photographs. To achieve this, we employ a novel mechanism based on the data field of an original image. Our inspiration is originated from physical fields. The used data field for an image can keep the balance between spatial and grayscale information in the local neighborhood by potential function, as well as the balance between the local information and the global trend. There involve three major steps in the proposed approach, including the generation of image data field, the reduction of potential centers, and the artistic renderings. The second step contains two types of reduction strategies, potential center elimination for stipple placement, and potential center cutting for mosaic construction. A number of experiments, both visual comparisons and quantitative comparisons, are performed. The results show the feasibility and the efficiency of the proposed method and suggest that the proposed method can generate appealing stipple or mosaic images. This would inspire graphic designers who may be interested in the stippled image automatically that is similar to images created by the artist.

Keywords:

Computational aesthetics

data field

digital stippling

halftoning

non-photorealistic rendering

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.