SVM-based synthetic fingerprint discrimination algorithm and quantitative optimization strategy.

Suhang Chen,Sheng Chang,Hao Wang,Qiangui Huang,Jin He,Qijun Huang,Rongrong Ji

doi:10.1371/journal.pone.0111099

Abstract

Synthetic fingerprints are a potential threat to automatic fingerprint identification systems (AFISs). In this paper, we propose an algorithm to discriminate synthetic fingerprints from real ones. First, four typical characteristic factors—the ridge distance features, global gray features, frequency feature and Harris Corner feature—are extracted. Then, a support vector machine (SVM) is used to distinguish synthetic fingerprints from real fingerprints. The experiments demonstrate that this method can achieve a recognition accuracy rate of over 98% for two discrete synthetic fingerprint databases as well as a mixed database. Furthermore, a performance factor that can evaluate the SVM's accuracy and efficiency is presented, and a quantitative optimization strategy is established for the first time. After the optimization of our synthetic fingerprint discrimination task, the polynomial kernel with a training sample proportion of 5% is the optimized value when the minimum accuracy requirement is 95%. The radial basis function (RBF) kernel with a training sample proportion of 15% is a more suitable choice when the minimum accuracy requirement is 98%.

Highlights

Fingerprint recognition is a relatively mature biometric identification [1,2] method, and automatic fingerprint identification systems (AFISs) have been widely used throughout our lives
Providing a fake fingerprint image is a simple and effective method of fraud that poses a considerable threat to the security of AFISs
The ridge distance average feature and ridge distance standard deviation feature, the global gray average feature and the global gray variance feature, the frequency feature and the Harris Corner feature are extracted from the image to constitute the feature vector, as defined in (1), in which InputMatrix is the input matrix of the support vector machine (SVM), Vm is the mth fingerprint image’s feature vector, and Qmn is the mth fingerprint image’s nth feature factor

Summary

Introduction

Fingerprint recognition is a relatively mature biometric identification [1,2] method, and automatic fingerprint identification systems (AFISs) have been widely used throughout our lives. Because AFISs normally connect with interests, attacks on such systems are ongoing. N.K. Ratha proposed a biometric system with eight possible attack points [3]. Providing a fake fingerprint image is a simple and effective method of fraud that poses a considerable threat to the security of AFISs. There are three typical fake fingerprints: altered fingerprints, non-living fingerprints and synthetic fingerprints. Altering a fingerprint directly changes the texture of the finger by obliteration, distortion and imitation. Because altering a fingerprint causes irreversible damage to the finger, it is not widely used except by criminals

Methods

Results

Conclusion