Weighted linde-buzo-gray stippling

Abstract

Fig. 1. Starting from a single random point, our dynamic illustration method creates high-quality stipple drawings with a small number of iterations (left to right: 12, 14, and 18 iterations). In the example shown above we end up with 36k points of constant size.We propose an adaptive version of Lloyd's optimization method that distributes points based on Voronoi diagrams. Our inspiration is the LindeBuzo-Gray-Algorithm in vector quantization, which dynamically splits Voronoi cells until a desired number of representative vectors is reached. We reformulate this algorithm by splitting and merging Voronoi cells based on their size, greyscale level, or variance of an underlying input image. The proposed method automatically adapts to various constraints and, in contrast to previous work, requires no good initial point distribution or prior knowledge about the final number of points. Compared to weighted Voronoi stippling the convergence rate is much higher and the spectral and spatial properties are superior. Further, because points are created based on local operations, coherent stipple animations can be produced. Our method is also able to produce good quality point sets in other fields, such as remeshing of geometry, based on local geometric features such as curvature.

Accurate Synthesis of Multi‐Class Disk Distributions

“…The purpose of point sampling is to dynamically generate a set of points that embodies specific properties. It can be used for stippling [MALI10,DSZ17,MAAI17], but also for rendering and texture synthesis. One common requirement is to match the spectral profile of blue noise, and consequently there have been many improvements to the state of the art in this area, such as improvements in computation [BWWM10, LNW∗10, CYC∗12], various properties [LWSF10, CGW∗13] and analysis [WW11].…”

A computational approach for spider web‐inspired fabrication of string art

“…For example, halftoning is the process of generating a pattern of binary pixels that creates the illusion of a continuous‐tone image . Closely related is stippling where dots are drawn, both digitally or physically, in such a way that it looks like a target photo. Another interesting work is the generation of a QR code resembling an input image .…”

Neural Networks for the Classification of Building Use From Street-View Imagery

“…To summarize, the amount of distributed points and their respective size depend on the tone, texture, and the local importance. We use the stippling algorithm of Deussen et al (2017) due to its ability to locally vary the degree of abstraction while avoiding visual artefacts such as distracting regularities within the point sets (Deussen et al 2000;Secord 2002). The algorithm uses the original image and its corresponding grayscale CAM as input.…”

“…Hence, it is directly correlated to the local importance. The algorithm of Deussen et al (2017) dynamically distributes points with given sizes to match the local tonal value of the input image. Using grayscale CAMs allows a direct linear mapping from importance [0;1] to point size [min;max].…”