基于多分类器决策融合的印鉴真伪鉴别方法

印鉴的自动识别对于实验银行印鉴数字化及计算机管理，实现银行对公业务的通存通兑有重大意义，符合票据法的要求，是银行业发展的需要，本文研究了模式识别领域中印鉴识别这一问题，提出了特征抽取和分类的新方法，基于差图像的成分统计特征抽取方法，发展了骨架图像的结构特征抽取方法，提出了特征点距离特征和笔划曲率特征内两种新的有效特征，本文提出了基于支持向量机（support vector machine,SVM)多分类器融合印鉴鉴别的方法，试验结果表明，本文的方法大大提高了印鉴的鉴别能力和可靠性，效果令人满意。

Multi-Classifier Decision Fusion Based Seal Imprint Verification Approach

Seal imprint verification has an important value in the implementation of digital signature in banks for saving and exchanging in the whole city. This paper researches this question from the view of pattern recognition and image processing and proposes new methods in feature extraction and classifier. The differences of image based statistical feature extraction method and the skeleton image based structural feature extraction technique are developed. Two kinds of effective feature, i.e., the key point distance feature and the stroke curvature feature, are suggested. Based on the above features, a support vector machine (SVM) based multi feature and multi classifier fusion is proposed. Experimental results indicate that the proposed methods can significantly improve the recognition accuracy and reliability compared to the previous seal imprint automatic verification approach.