Research And System Design On Contactless Palmprint And Palmvein Recognition

Internet is changing people's daily life.It has deeply influenced the style of identity authentication.The worldwide deteriorating security situation requires a higher accuracy identity authentication method.It is a good opportunity for palmprint recognition system.When sampling with traditional contact palmprint recognition system,users' hands directly contact with the acquisition equipment.It is not convenient for users.However,current non-contact palmprint recognition systems suffer from complex algorithm,high time cost and low security performance.To solve the above problems,a new non-contact palmprint and palm vein multimodal recognition system is designed in this paper.To solve the problems of non-contact palmprint recognition,a new acquisition device is designed.Active light source and narrow bandpass filter are used in this device.They help to reduce the influence of ambient light.In that case,a simple center block localization algorithm can reach a good performance.In addition,the filter stabilizes the light intensity of sample images.Scale differences and defocus blur can be avoided by distance sensor,instead of image scale correction algorithm.The acquisition time is shortened by the dual-camera system,because palmprint image and palm vein image can be caught at the same time.In that way,the proposed acquisition device helps to reduce the algorithm complexity.The experiment results show that the acquisition device scheme is effective.The robustness to dislocation samples and the accuracy of living detection are essential for a non-contact palmprint identity authentication system.For palm vein recognition,a BRIEF algorithm based on the best matching region is proposed in this paper.The best matching region of two samples is obtained in sliding window,then the similarity of the two samples is measured by BRIEF feature.Because of the robustness to dislocation samples,this algorithm reaches an excellent performance.For palmprint recognition,a Gabor histogram algorithm based on subregions is designed.The algorithm provides ambiguity spatial information with a certain ambiguity and details information of texture as well.The experiment results show the effectiveness of this algorithm.In addition,a fusion recognition algorithm and three authentication strategies are designed to meet the needs of different occasions.