Silhouette Vectorization by Affine Scale-Space

Silhouettes are building elements of logos, graphic symbols and fonts. These shapes can be designed exchanged in vector form, but more often they drawn, printed, scanned, or directly found digital images. Such raster forms require vectorization to get scale-invariant exchangeable formats. There is a need for mathematically well-defined justified shape process, which also provides minimal set control points with geometric meaning. In this paper, we propose new silhouette paradigm. It extracts the outline 2D from binary image converts it combination cubic Bézier polygons perfect circles. The proposed method uses sub-pixel curvature extrema affine scale-space vectorization. By construction, our geometrically stable under transformations. used as reliable feature point detector silhouettes. Compared state-of-the-art software, algorithm demonstrates superior reduction number while maintaining high accuracy.