Online Handwriting Mongolia Words Recognition Based on HMM Classifier

Wu Wei,Gao Guanglai
DOI: https://doi.org/10.1109/ccdc.2009.5191525
2009-01-01
Abstract:In this paper, we give a recognition method for Online Handwriting Mongolia words which being often used among the Mongolia people in the North China. We introduced the HMM technique for handwriting recognition. We make use of local dynamic information for feature selection, and k-means clustering algorithm base on HMM. Experimental results show that writer-dependent words achieve recognition rates above 90%. The system runs well and is good stability. Recognition rate achieves just to the level of utility.
What problem does this paper attempt to address?