Discrete Wavelet Transform Based Approach for Touchless Fingerprint Recognition

Abstract

AbstractThe fingerprint-based human recognition is very much popular and commonly used approach. Currently, employed fingerprint-based biometrics system is contact- or touch-based. Touch-based systems have problems like elastic deformation, hygiene, maintenance, low contrast samples. One alternative to this is touchless fingerprint recognition. Due to current situation of Covid-19 outbreak touchless recognition is very much safe. Several researchers have carried out work in this area but still it needs improvement in performance. In contactless fingerprint, recognition fingerprint images are captured using cameras without any direct contact. Existing touch-based fingerprint systems are limited in performance due to variations in contrast, illumination and magnification. The current touch-based fingerprint system is having less efficiency in extraction of reliable minutiae. Proposed research work presents a touchless fingerprint recognition system. The proposed touchless fingerprint recognition system uses fingerprint images acquired from digital camera. Proposed system employs discrete wavelet transform-based feature extraction algorithm. Fingerphoto features are extracted by applying wavelet decomposition using 6 different wavelet families and 26 respective wavelet family members. Hamming distance-based classifier has been used for matching of wavelet transform-based features. Proposed touchless fingerprint identification system has been evaluated on Hong Kong Polytechnic University database. We have achieved better results for our proposed system as FAR = 0.27, FRR = 3.91 with Genuine Accept Rate of 96.09.KeywordsTouchlessFingerprintWaveletBiometricsIdentification