SIGGRAPH 2021: StrokeStrip will cleanup your drawings for you

We continue our dive into the SIGGRAPH 2021 technical papers with a fascinating project aimed at sketching and vector drawing: StrokeStrip : joint parameterization and fitting of stroke clusters.

This technical paper was written by Dave Pagurek van Mossel in collaboration with Chenxi Liu (University of British Columbia), Nicholas Vining (University of British Columbia & NVIDIA), Mikhail Bessmeltsev (University of Montreal), Alla Sheffer (University of British Columbia).

When creating sketches, artists use multiples strokes instead of just one curve. A new stroke can be added by the artist to refine the shape already created, or to add thickness. The intended drawing is easy to guess for the human eye:

Such a task, however, is far from obvious for a computer. Existing methods (Orbay et Kara – 20211, Lieu et al. – 2018) are far from perfect. They produce interesting results but lack robustness. The authors present StrokeStrip, a new method that gives much better results. StrokeStrip relies on a clever parametrization of each of the strips created by the strokes.

Moreover, the parametrization used by StrokeStrip can have other useful applications, such as sketch editing or coloring.

For more information, please check out the official webpage for this project: StrokeStrip : joint parameterization and fitting of stroke clusters. Note that a C++ implementation of the algorithm is provided. For a technical explanation of the proposed method, you can also watch the following videos by Dave Pagurek. The paper is explained in 30 seconds, 5 minutes and 20 minutes: feel free to chose the video that fit your needs, whether you just want a short explanation or a deep dive.

A Lire également

Laissez un commentaire