3D reconstruction from planar points: A candidate method for authentication of fingerprint images captured by mobile devices

摘要

With the widely application of mobile commerce (m-commerce), there is an urgent need to introduce identity authentication mechanism to mobile devices and make m-commerce secure to users. Since fingerprints can be viewed as an important characteristic of a natural person, fingerprint authentication (FA) can provide excellent access control for mobile phone users. Generally speaking, there exist two methods for applying of FA to mobile devices: one solution is trying to introduce an extra sensor to a mobile device, the other solution is trying to authenticate an image captured by a mobile device directly. However, the drawback of the first method is that devices with specific sensors are usually unavailable, the critical difficulty for the second solution comes from minutiae matching since different images would be obtained by a camera with different snap in different positions are different. Facing with these difficulties, this paper trying to provide a possible novel method which combines 3D reconstruction with minutiae matching: reconstruction of 3D minutiae from two 2D images first and then match the obtained image with the reconstructed 3D minutiae.