Mimicking Hand-Drawn Pencil Lines

Abstract

In applications such as architecture, early design sketches often mislead the target audience [SSRL96]. Approximate human-drawn sketches are typically accepted as a better way of demonstrating fundamental design concepts. To this end we have designed an algorithm that creates lines that perceptually resemble human-drawn lines. Our algorithm works directly with input point data and physically based mathematical model of human arm movement. Further, the algorithm does not rely on a database of human drawn lines, nor does it require any input other than the end points of the lines to generate a line of arbitrary length. The algorithm will generate any number of aesthetically pleasing and natural looking lines, where each one is unique. The algorithm was designed by conducting various user studies on human line sketches, and analyzing the lines to produce basic heuristics. We found that an observational analysis of human lines made a bigger impact on the algorithm than a statistical analysis. A further study has shown that the algorithm produces lines that are perceptually indistinguishable from straight hand-drawn pencil lines.