Title :
Ribbon-like skeletonization based on contour reconstrction on intersection regions
Author :
Liao, Melody Z W ; Huang, X.Q. ; Hu, S.X. ; Chen, W.F.
Author_Institution :
Sch. of Comput. Sci., Sichuan Normal Univ., Chengdu, China
Abstract :
According to the requirments of ¿good¿ skeletons, we present a skeletonization based on repairing the contours, especially for intersection regions. The motivation of this method is each of Chinese character is composed by some strokes, and the ¿good¿ skeletons can be extracted easily from these strokes using existing methods. Therefore, the contours on intersection regions can be reconstructed correctly by repairing strokes. In this paper, the contour lines and corner points will be detected firstly; and then the strokes are recovered by repairing the contour lines detected on step 1 according to convex and concave natures of distinct points; Finally, the skeletons are extracted using strokes. Experimantal results show the good performance of our method.
Keywords :
character recognition; edge detection; image reconstruction; image thinning; wavelet transforms; Chinese character; contour reconstruction; corner point; intersection region; ribbon-like skeletonization; stroke extraction; wavelet maxima; Automation; Character recognition; Computer science; Cybernetics; Data mining; Humans; Noise reduction; Noise robustness; Skeleton; USA Councils; Skeletons; contour; contour repaired; skeletonization; stroke;
Conference_Titel :
Systems, Man and Cybernetics, 2009. SMC 2009. IEEE International Conference on
Conference_Location :
San Antonio, TX
Print_ISBN :
978-1-4244-2793-2
Electronic_ISBN :
1062-922X
DOI :
10.1109/ICSMC.2009.5346052