Shape Manipulation of Diffusion Curves Images
Published in IEEE Access, 2020
Abstract: Vectorization converts raster scans of line drawings into vector graphics; it breaks the barrier between line drawing generation and postprocessing. Prior work on line drawing vectorization considerably succeeded in revealing artists’ drawing intention driven by structural topologies. However, none of them is able to extract simplified topologies for sketchy line drawings consisted by many unwanted lines. In this paper, we propose an improved topology extraction approach based on artists’ sketching customs. Redundant regions and open curves are discriminated from artists’ deliberate ones and further removed progressively through an iterative optimization mechanism. We demonstrate that our improved topology benefits our vectorization method as well as existing topology-driven ones and allows them to vectorize rough sketchy line drawings robustly and efficiently.
Recommended citation: Shufang Lu, Xuefeng Ding, Fei Gao, Jiazhou Chen. “Shape Manipulation of Diffusion Curves Images.” IEEE Access. 2020, 8: 57158-57167.