欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对单模态生物特征识别的局限性,提出融合手背静脉、虹膜和指纹三种生物特征实现身份识别。首先分别对手背静脉图像、虹膜图像和指纹图像进行独立的图像预处理、特征提取和特征匹配,输出各自的匹配分数。分析匹配分数归一化对识别性能的影响,采用Tanh归一化方法对三种生物特征的匹配分数进行归一化处理,最后利用加权求和法则实现匹配分数的融合,利用最小距离分类器实现身份识别。实验结果表明,融合识别算法的等错率为0.009%,当错误接受率接近0时,对应的错误拒绝率仅为0.2%。
英文摘要：
Because of the limitation of single biometric recognition,a new multimodal biometric system based on hand vein,iris and fingerprint was proposed.Firstly,the hand vein image,iris image and fingerprint image were separately sent to image preprocessing module,feature extraction module and feature matching module,respectively.And the respective matching scores were obtained.The effects of score normalization on recognition performance were analyzed.Tanh normalization method was employed to perform the three scores normalization.Finally,the weighted sum principle was used to fulfill the scores fusion.Minimum distance classifier was adopted to personal identity decision.The experiments results show that Equal Error Rate（EER） of proposed recognition algorithm is 0.009%.When the False Acceptance Rate（FAR） is close to 0,the corresponding False Rejection Rate（FRR） at the same threshold is 0.2%.
同期刊论文项目
同项目期刊论文
期刊信息
位置：