ScholarWorks@성균관대학교

초록

Binary black-and-white sketches are widely used in fields such as clothing design due to their clarity and low storage requirements. Despite these advantages, the inherent inflexibility of raster images makes them inefficient in redesign tasks, as modifying or scaling the sketch requires a conversion to an editable vector format. To address this limitation, we propose a method for converting binary sketches into Scalable Vector Graphics (SVG), as this process serves as a crucial intermediate phase for consumer product customization and design modifications. Our proposed method involves raster preprocessing and vectorization, followed by topology-preserving point minimization and final curve refinement to maintain contour fidelity. The performance was assessed in terms of point reduction ratio and accuracy. Experimental results show the effectiveness of our proposed method on our binary sketch samples. The number of vector points was substantially reduced - from thousands to a few hundred - resulting in an average point-reduction ratio of 89.52%, while maintaining an average structural accuracy of 98.86%.

키워드