Handwritten Digit Recognition by Using Bayes Classification
This paper proposes a novel Bayes classification algorithm based on Minimum-risk decisions to recognize handwritten digits. One existed algorithm uses the classic conditional probability in the Bayes model to construct a set of Risk functions and thus estimates the most possible digit by making decisions based on Minimum-error. In this process,a series of probability distribution functions describe the features of images. To improve the recognition rates further,the proposed model adjusts weights of the Risk functions to make decisions based on Minimumrisk. Experimental results show that the proposed method has a higher recognition rate than the original one and other popular recognition methods.
Pattern recognition Handwritten digit recognition Bayesian classification Minimum-error Minimum-risk
Heng Li Xinru Wang
Department of Electronic Engineering,Fudan University,Shanghai 200433,China School of Computer Science,Chongqing University,Chongqing 400044,China
国际会议
西安
英文
1297-1300
2011-12-23（万方平台首次上网日期，不代表论文的发表时间）