Multimodal Hand Feature Fusion Recognition

Single modal biometrics,such as palm-print recognition,iris recognition,gesture recognition and so on,have gained more and more widespread applications,but due to the inherent limitations of their single mode,they can not satisfy the current social security Claim.The multi-modal fusion technology is the same modality of a number of different biological characteristics,by adopting some kind of fusion rules,fusion as a whole to identify,with strong security and identifiability.Although multi-modal fusion technology has become a research hotspot in recent years,there are still many problems waiting to be solved.In this paper,based on the fusion level of multi-modal fusion information,the fusion algorithm based on score layer fusion and feature layer fusion is studied.The main contents are as follows:(1)Adaptive multimodal biometrics fusion based on classification distance score Matching score is one of the traditional fusion score metrics,but it's not a good metric to classify the data with intra-class and inter-class scores.The classification confidence score can be used to well separate the data with intra-class score from the data with inter-class score,but it does not work well for the data whose matching scores are next to the classification threshold.Therefore,this paper proposes a new score metric based on the classification distance score,which contains not only the information of the first level of classification,but also the information of the distance between matching score and classification threshold,and which can also increase the distance of the fusion scores between intra-class and inter-class scores,and the classification distance score provides the characteristics of effective discriminative information fusion set for fusion algorithm,which can improve the utilization rate of score metric;furthermore,since the information entropy indicates the information value of features,it can be used to define the feature correlation coefficient and feature weight coefficient,and then,the weighted fusion and traditional SUM rules are unified in an adaptive algorithm framework,which can improve the fusion recognition rate.The experimental results indicated the validity of the proposed method.(2)Multimodal biometrics fusion based on label discriminate correlation analysisCanonical correlation analysis of the traditional method and its improvement through restraint the method of within class scatter matrix and inter class scatter matrix,which achieves the purpose of using the class label information indirectly,and not directly integrate the class label information with the feature set,which reduces the application of fusion algorithm in a certain degree on.Therefore,first proposes the algorithm to fuse the feature set and the class label information,establishes criterion function for it,obtains the optimal projection vector by using Lagrange function,and then acquires feature sets with the class information,which can provide the characteristics of more discriminating information feature sets for the subsequent algorithm;further,for feature sets with the class information,minimize the within class scatter matrix,maximize the covariance matrix between the feature sets of two modes at the same time,in order to extract the feature vectors with more discriminating information,which has higher ability to identify and improve the recognition rate.The experimental results verify the effectiveness of the proposed method and the rationality of the combination of the two algorithms.