Method of 3D biometrics identification

摘要

The present invention discloses a contactless 3D biometric feature identification system and the method thereof. The system comprises of a fixed-viewpoint image capturing means, a lighting module capable of producing different illuminations on the object of interest and a microprocessor configured to execute a biometric identification algorithm. The algorithm starts with capturing a plurality of images with different illuminations. The captured images are then utilized to reconstruct a three dimensional surface model. Different features, for instance 2D and 3D coordinates and orientations of the biometric feature, surface curvature of the object and the local surface orientation of the object, are extracted from the captured images and the reconstructed 3D surface model. Different matching scores are also developed based on the aforesaid features to establish the identity of the biometric features.

主权项

1. A method executed by a computer system to identify a fingerprint of a finger of a user, comprising:
capturing, by the computer system, a plurality of 2D images of the fingerprint of the finger of the user; reconstructing, by the computer system, a 3D surface model of the finger from the 2D images; preprocessing, by the computer system, the 3D surface model through a median filter; extracting, by the computer system, a plurality of 3D minutiae from the 3D surface model; computing, by the computer system, a difference between a 3D minutia of the finger and a template minutia based on heights, locations, and angles by:funRSG⁡(Δ⁢⁢v)=(Δ⁢⁢r65)0.8+(Δ⁢⁢As30)0.8+(Δ⁢⁢Ag15)0.8+(Δ⁢⁢Aθ18)0.8+(Δ⁢⁢Aφ42)0.8+(1-cos⁢⁢N0.075)0.8, where Δr is a difference between the radial distance of the 3D minutia of the finger and the radial distance of the template minutia, ΔAθ is a difference between the azimuth angle of the 3D minutia of the finger and the azimuth angle of the template minutia, ΔAφ is a difference between the elevation angle of the 3D minutia of the finger and the elevation angle of the template minutia, ΔAs is a difference between the azimuth angle of the radial vector of the 3D minutia of the finger and the azimuth angle of the radial vector of the template minutia, ΔAg is a difference between the elevation angle of the radial vector of the 3D minutia of the finger and the elevation angle of the radial vector of the template minutia, and N is the normal of the 3D minutia of the finger; determining, by the computer system, whether the 3D minutia of the finger matches the template minutia; generating, by the computer system, a matching score based on a number of matched minutiae pairs; and establishing, by the computer system and based on the matching score, an identity of the user.