关键词: 联机手写识别, 维吾尔文字母, 时间弯折算法, 多分类器集成, 投票法

Abstract: Combination of multiple classifiers, a certain extent, compensate for defects of a single classifier, so it has been widely applied in pattern recognition. Through in-depth study of the trend of on-line handwriting recognition technology in domestic and abroad, based on analysis of the unique shape of Uyghur characters, this paper studies the Uighur online handwritten character recognition based on multi-classifier combination. It applies five different feature extraction methods to construct five separate classifiers of Uyghuer characters and using voting strategy of ranging from rights to effective combined five kinds of classifier of Uyghuer characters. Each classifier uses the nearest neighbor classification method based on Dynamic Time Warping（DTW） matching distance. Experimental results show that the recognition rate based on integration strategy is significantly higher than the recognition rate of separate classifier, and it also provides a variety of effective ways for the comprehensive integration of features.

Key words: online handwriting recognition, Uyghur characters, Dynamic Time Warping（DTW）, multi-classifier combination, voting