Research On 3D Fingerprint Recognition Technology Based On The Space Feature Points

Nowadays,the research and application of fingerprint identification has been extended to all areas of life.The traditional acquisition of fingerprint by pressing or rolling of the finger against the hard surface often results in partial or degraded quality images due to improper finger placement,skin deformation,slippages,smearing or sensor noise,which will affect the accuracy of fingerprint identification.However,finger itself is three dimensional,containing the breadth of the 3D information.Because of the acquired depth information of the fingerprint,3D fingerprint identification technology can overcome a series of problems in the traditional 2D fingerprint recognition technology,and it also has a good robustness for illumination,stains and other issues.This paper presents a 3D fingerprint identification technology based on space feature points.We attempt to expand the traditional 2D feature points to 3D space by increasing the depth information and direction information and extract 3D feature points of fingerprint.3D fingerprint data are obtained by using a fast 3D fingerprint acquisition system based on fringe projection.The collected 3D fingerprints are mapped to a 2D plane.The feature points of fingerprint have been pre-processed.The feature points' location information is record and mapped to the 3D space to acquire location coordinates of the space feature points.The 3D direction information is calculated according to the coordinates of feature point,i.e.the space characteristics of 3D fingerprint feature points are extracted absolutely.The space feature points extracted from template fingerprint and matched fingerprint are aligned based on double reference point method.The aligned feature points are matched according to the euclidean distance between the feature points and the direction angle.Then the matching score is calculated to measure the similarity of two fingerprints.When the matching score exceeds the setting threshold value,argues that the two fingerprints from the same finger,otherwise matched failure.The experimental results show that the equal error rate can reach 1.32%,however the EER of traditional 2D fingerprint identification system is difficult to reduce to less than 2%,proving that based on the space feature points of 3D fingerprint matching recognition method is effective and feasible.This method not only obtains the 3D fingerprint characteristic information to a great extent,improving the recognition accuracy effectively,but also provides reliable method and technical route for the development of 3D fingerprint identification technology,and expands its theoretical research.