Research On3D Reconstruction Based3D Fingerprint Recognition Technology

3D reconstruction based touchless fingerprint recognition is a newly proposedapproach working for biometric identification/authentication. The touchless fingerprintimages can help solve the problems suffered in traditional touch-based fingerprintsystems, including skin deformation, the latent fingerprints, etc. Furthermore, the "real"3D fingerprint model could be built based on several images captured from differentviews using3D reconstruction. Within such3D models, more discriminative featurescould be extracted and more accurate identification/authentication results could beobtained.3D fingerprint recognition has become a research hotspot due to its advantages andthe growing demand for biometric identification/authentication. This thesis is focusedon some issues related to3D fingerprint recognition, including touchless fingerprintimage pre-processing，3D fingerprint model reconstruction and3D fingerprint matching.The main contributions can be exhibited by the following aspects:1. A new3D fingerprint image pre-processing method is proposed aiming atsolving several problems existed in camera based fingerprint image, including distortion,complex background, low contrast, etc. Integrating image correction, segmentation andenhanced functions, the new pre-processing method could eliminate image distortionand background noise,and could enhance the contrast of fingerprint ridge. Experimentalresults show that the proposed method is effective. The result images afterpre-processing are more compatible with the traditional2D fingerprint processingmethods.2. A ridge based3D fingerprint model reconstruction method is proposed. Theridge pattern of fingerprint in3D space could be constructed based on some existed2Dfingerprint processing methods and the mathematical model of multiple views.Compared with other finger surface reconstruction methods, the data amount of themodel obtained by the proposed method is smaller. The obtained model is also expectedto be more suitable in the subsequent steps, such as feature extraction, classification,etc.3. A3D minutia based3D fingerprint matching method is proposed. Afterextending the minutia based2D fingerprint matching method to3D space, based onsome existed2D fingerprint processing methods and the mathematical model ofmultiple views, the3D minutia could be accurately extracted and matched.Experimental results show the effectiveness of the proposed method.