100 likes | 123 Views
Divide&Conquer based Vectorization. Mgr. Michal R é cky Comenius University, Bratislava. Image vectorization. Raster image transformed to vector image First step in object recognition Edges are transformed into vectors. Divide&Conquer approach.
E N D
Divide&Conquer based Vectorization Mgr. Michal Récky Comenius University, Bratislava
Image vectorization • Raster image transformed to vector image • First step in object recognition • Edges are transformed into vectors
Divide&Conquer approach • Idea: vectorization can be solved easily in small segments of image • Fast and relative simple algorithm compared to other methods • Can be applied on complex (real) images
Divide • Edge detector is applied on image • Image is divided into cells • Vectorization problem is solved inside these cells separately
Types of cells • Simple line • Arc • T-junction • Complex junction
Conquer • Set of vector is transformed into graph • Joining close vertexes • Removing vertexes at straight lines
Summary • More testing is required • Extensive practical application • Divide&conquer based vectorization has the potential to become the most effective method available
Thank you for your attention Mgr. Michal Récky recky@fmph.uniba.sk