基于数据融合的多生物特征身份识别

Biometrics Personal Identification Based on Dada Fusion

针对传统的单一生物特征身份识别由于传感器的噪音以及特征的破损、匹配等缺陷,往往不能正确识别的情况,提出了一种基于语音和人脸的多生物特征身份识别方法.分别提取语音特征和人脸特征作为识别的依据,并用神经网络在特征层上进行融合识别.实验证明,这种方法可以充分挖掘特征之间的关系,在同等条件下,具有更高的识别率.

Personal identification is one of the most important fields on pattern recognition. Duo to sensor noise, missing of characters or the bugs of matching algorithm, conventional recognition systems based on single-character can not give a true result. The paper presents an intelligent recognition system of multi-character fusion, which is acoustic and face based on artificial neural network. The experiment is shown to be superior to that of acoustic and visual subsystems.