Title :
An algorithm of handwritten digits segmentation based on multi-mould
Author :
Zhang, Hong-Gang ; Liu, Gang ; Xu, Wi-Ran ; Guo, Jun
Author_Institution :
Inf. Eng. Sch., Beijing Univ. of Posts & Telecommun., China
Abstract :
An algorithm of cursive handwritten digits segmentation based on multi-mould is proposed. The algorithm summarizes the characteristics between touching handwritten digits, and sums up a serial of curves. Based on these curves, we design several kinds of segmentation moulds, which change the segmentation process into probation and select the best of all kinds of moulds. The application in bank cheque OCR system verified the validation of this algorithm.
Keywords :
bank data processing; handwritten character recognition; image segmentation; optical character recognition; bank cheque OCR system; cursive handwritten digits segmentation; handwritten character recognition; multiple mould method; segmentation path; serial curves; touching handwritten digits; Algorithm design and analysis; Cybernetics; Dynamic programming; Heuristic algorithms; Mathematical model; Mathematics; Optical character recognition software; Probes; Shape; Viterbi algorithm;
Conference_Titel :
Machine Learning and Cybernetics, 2002. Proceedings. 2002 International Conference on
Print_ISBN :
0-7803-7508-4
DOI :
10.1109/ICMLC.2002.1174550
