Font Size: a A A

Algorithm Research Of Multimodal Biometric Identification Technology

Posted on:2018-10-31Degree:MasterType:Thesis
Country:ChinaCandidate:Y ZhangFull Text:PDF
GTID:2348330536459572Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
Recently,the biometric identification technology has developed fast,but the traditional single-mode identification system has so many limitations taht result in practical application of this technology in the many difficulties.With the emergence of data fusion technology,multimodal biometric identification that uses a variety of biometric features for data fusion identification technology got a lot of technical support,and this technology can be faster into our daily lives.In this paper,through the study of the characteristics of common biological features and the fusion methods of multi model biological characteristics,two biometric fingerprint and iris were used,and in multimodal biometric features in the feature layer fusion of biometric identification experiment was carried out.The main contributions are as follows:1.Study on the method of biological recognition,the iris recognition and fingerprint recognition technology of two kinds of biological characteristics,especially in view of the above two kinds of single biological feature extraction techniques are studied and improved in some application methods.2.Study on the related methods in the field of multimodal biometrics,including series parallel and based on Canonical Correlation Analysis(CCA)fusion,least square method and the Fisher discriminant score level fusion,majority voting and weighted method in decision fusion layer based on feature layer.3.Based on defect of canonical correlation analysis,a Matrix transform Discriminant Canonical Correlation Analysis(MDCCA)based on the multi modal biometric recognition algorithm was proposed,the feature vector extraction and processing good iris and fingerprint in feature level feature fusion,finally,recognition experiments verify the effectiveness of the algorithm.4.Based on the new algorithm proposed in this paper,a set of multi modal biometric identification system is constructed.The research content of this paper has a certain reference value in the field of single modal identification,and it has a certain reference value in the field of multimodal biometric identification.
Keywords/Search Tags:Multi-mode, Fusion recognition, Iris identification, Fingerprint identification, MDCCA
PDF Full Text Request
Related items