Palmprint And Palm Vein Fusion Recognition Algorithm And System Implementation

As the development of information technology,the idea of protecting self information is stronger and stronger.Traditional technology of identity authentication such as magnetic card,password etc.is no longer safe.On the other hand,biometric authentication technology is high security and easy to use,which has been accepted and widely used in many areas.One of the topic researchers is biometric recognition of hands,which includes finger vein,fingerprint,palm vein and palm print recognition.Note that single feature of palm is simple so that its recognition rate is not high enough.Researchers proposed a method named multi-features fusion recognition of hands and the fusion recognition of palm vein with print is one of the research hotspots.Existing methods of fusion recognition of palm vein with print are almost based on two independent images that are hard to acquire and to ensure the coherence of corresponding areas.In fact,single near-infrared palm image is comprised of not only the structure of palm vein,but also the information of palm prints that can be used to fusion recognition.This paper aims to research the method of fuse the palm vein with palm prints in a single near-infrared palm image.The main contributions of this paper are summarized as follows.Based on the texture structure of palm vein,we proposed an enhanced algorith of palm vein that is based on Weighted Guided Filtering for Edge Detection.First of all,we employ the self-guided filtering methed to preprocess the near-infrared palm image twice.The first can preserve the structure information of palm vein by using a smaller filter radius,whereas the second aim to wipe off the structure information of palm prints by using a larger filter radius.Secondly,a linear enhancement model was designed to enhance the palm vein structure.Finally,the above steps can be repeated to obtain an enhanced image of the palm vein structure.The algorithm adaptively processes the vein edge and smooth region in the image while enhancing the structure of the palm vein,so as to better protect the edge structure information of the palm vein.The experimental results show that the proposed algorithm can effectively extract the palm vein structure information,and the single recognition rate of the palm vein structure reaches 99.28%.Based on the characteristics of palm print structure,a fuzzy anti-sharp mask palm print enhancement algorithm based on block model was proposed.The main ideas are as follows:Firstly,the block model algorithm is used to remove the palm vein structure information in the near-infrared image.Secondly,the image after removing the vein structure information was blurred using the membership function.Then,the fuzzy theory and the anti-sharp mask are masked.The membrane algorithm was combined to obtain the enhanced palm print matrix,and the enhanced palm print matrix was normalized to[0,255],which is the enhanced palm print structure.The experimental results show that the proposed algorithm can effectively extract palm print structure information,and the single recognition rate of palm print structure reaches 95.68%.Based on the above palm vein and palm print enhancement algorithm,a palm vein palm print fusion recognition system was designed and implemented.It mainly includes palm entry,palm image authentication,password authentication and other functions.