摘要
提出了一种新型的笔迹鉴别方法.它通过局部特征匹配与投票来提取关键词并建立训练集,对两篇文档的相同关键词进行匹配,将笔迹鉴别问题转化为签名识别问题.实验表明,在文档内容部分相同的情况下,本方法能够有效地对文档进行笔迹鉴别.
In text independent hand textures and other handwriting feat writing verification ures; while in text , lengthy documents are required to extract image dependent handwriting verification, the contents of the compared documents are required to be identical. In order to overcome these two confinements, this paper presents a novel text dependent handwriting verification method. In the process, keywords are extracted based on matching and voting of local features, then the same spelling keywords are used to build training sets, and these train sets of two documents are compared, so that the handwriting verification problem is transformed into a signature identification problem. The experiment results show that this method can work effectively under the condition that the contents of two documents are partially identical.
出处
《西南大学学报（自然科学版）》
CAS
CSCD
北大核心
2014年第1期137-145,共9页
Journal of Southwest University(Natural Science Edition)
基金
中央高校基本科研业务费专项资金资助项目(XDJK2012C068)
教育部春晖计划项目(Z2011149)
关键词
局部特征
匹配与投票
参考表
笔迹鉴别
签名识别
local feature
matching and voting
reference table
handwriting verification
signature identification
