摘要
计算机印鉴自动真伪鉴别是一项十分困难、但又具有广泛应用前景的一项综合性研究课题．国内外至今尚缺乏实用有效的方法．我们从图象分析、模式识别原理出发，给出了一套完整的印鉴鉴别方法．该方法充分提取了反映真假印鉴本质特征的多种特征参数，结合了统计模式识别与结构模式识别方法各自的长处，使得鉴别方法在准确性、稳定性、可靠性与适应性方面与我们原有的方法以及国外其他方法相比，都取得了很大的进展，已达到应用水平．
Based on the principle of image analysis and pattern recognition,this presentation proposes an improved processing algorithm for the seal imprint verification. As the approach makes full use of the multiple intrinsic features of the seal imprints,and uses the statistical and structural pattern recognition methods complemently, it greatly improves the accuracy , stability,reliability and compatibility of its performance, comparing to our old method and other already proposed approaches,and can be put into practical use.
出处
《计算机学报》
EI
CSCD
北大核心
1995年第3期190-198,共9页
Chinese Journal of Computers
基金
中国工商银行科技部研究基金
关键词
印鉴
鉴别
模式识别
图象处理
Computer image processing, pattern recognition, feature extraction,nonlinear classifier.
