中文摘要：

单一生物特征识别技术无论是在识别率还是稳定性上都不能达到完美无缺,特别是高仿生物特征的出现使其安全性受到质疑。针对上述问题,提出一种手形、掌纹和掌静脉多特征融合识别方法。提出了基于小波变换Gabor滤波器的特征层和图像层融合策略,将同一设备不同光照采集下的掌纹和掌静脉融合到一起,突出各自的主纹理特征;利用手指相对长度为手形特征进行初匹配,提出利用分块纹理基元模型进行掌纹和掌静脉融合图像的特征提取方法,然后进行二次匹配给出最终识别结果。开发了模拟系统并进行了相应的实验,结果表明该识别系统充分发挥了3种特征各自的优点,提高了识别率和稳定性,特别是掌静脉的加入增强了系统的安全性。

英文摘要：

Unimodal biometric identification technology cannot achieve perfect recognition rate and stability,especially the emergence of the imitation of biological characteristics make its security being questioned. Aiming at this problem,in this paper,a multimodal feature fusion identification method based on hand shape,palmprint and palm vein is presented. The fusion strategy of feature extraction level and image level is proposed based on Wavelet transform and Gabor filtering,which fuses the palmprint and palm vein images captured under different light conditions by the same acquisition equipment. The main texture information of the fused image is emphasized. The hand finger relative length is used as the hand shape characteristic to conduct the first-matching. A feature extraction algorithm of the palmprint and palm vein images that applies block？ texture primitive based on statistics is proposed. Then,the second matching is performed to obtain the final recognition result. The simulation system was developed and the proposed algorithm was tested using the developed system. Experiment results show that the proposed recognition system makes full use of the respective advantages of three modal characteristics,improves the recognition rate and stability; especially the introduction of palm vein enhances the security of the system.