Finger Vein Recognition Based On Local Features

In recent years,with the rapid development of information technology,people pay more attention to the security of information.Biometric uses the inherent biological characteristics of human for identity recognition,and is not easy to be forgotten and stolen,therefore it has a wide range of application in information security.As a kind of liviness biometrics,finger vein has the merits of stability and easy collection,therefore,it has some advantages compared with other biometrics.However,the finger vein image is easy to be influenced by external factors such as temperature,humidity,illumination,etc.Therefore,it has certain difficulties during the identification.And traditional methods have some limitations,in order to make full use of rich directional information and spatial information of finger vein,this thesis proposes some kinds of finger vein recognition methods based on local features.The main works of this thesis are shown below.First,this thesis carries out the research on the background and significance of the finger vein recognition technology and the advantages of making use of local features for finger vein recognition.Second,making full use of the rich directional information and spatial information of finger vein,this thesis proposes three kinds of operators,which are Multi-Orientation Weighted Symmetric Local Graph Structure,Difference Symmetric Local Graph Structure and the Cycle Gradient Operator,and applies them to the finger vein recognition.The three operators are compared with the existing classical operators.The experimental results show that all the three operators improve the efficiency of the finger vein recognition algorithms,and have a certain advantage in the recognition rate.