Abstract
The performance of conventional minutiae-based fingerprint authentication algorithms degrades significantly when dealing with low quality fingerprints with lots of cuts or scratches. A similar degradation of the minutiae-based algorithms is observed when small overlapping areas appear because of the quite narrow width of the sensors. Based on the detection of minutiae, Scale Invariant Feature Transformation (SIFT) descriptors are employed to fulfill verification tasks in the above difficult scenarios. However, the original SIFT algorithm is not suitable for fingerprint because of: (1) the similar patterns of parallel ridges; and (2) high computational resource consumption. To enhance the efficiency and effectiveness of the algorithm for fingerprint verification, we propose a SIFT-based Minutia Descriptor (SMD) to improve the SIFT algorithm through image processing, descriptor extraction and matcher. A two-step fast matcher, named improved All Descriptor-Pair Matching (iADM), is also proposed to implement the 1:N verifications in real-time. Fingerprint Identification using SMD and iADM (FISiA) achieved a significant improvement with respect to accuracy in representative databases compared with the conventional minutiae-based method. The speed of FISiA also can meet real-time requirements.
Highlights
Fingerprints have been used as a personal identification tool for a long time because of their uniqueness and time invariance
Using the proposed Scale Invariant Feature Transformation (SIFT)-based minutia descriptor (SMD), we developed a two-step fast matching method, called improved All Descriptor-Pair Matching
In order to demonstrate the performance of Fingerprint Identification using SMD and iADM (FISiA), several well-known fingerprint databases were tested
Summary
Fingerprints have been used as a personal identification tool for a long time because of their uniqueness and time invariance. The performance of this system will degrade significantly if the overlapping area between the template and the input fingerprint image is small, and when the number of available minutiae is few. This case occurs when a large translation of finger position occurs or when a swipe sensor with a very narrow width is used. This paper proposes a new fingerprint verification algorithm using SIFT-based minutiae descriptor (SMD). The minutia descriptor proposed in this paper employs SIFT [10,11] information as the additional data in order to reduce the complexity of feature correspondence. We present the flow of image processing, minutiae extraction and the definition of SIFT-based Minutia Descriptor
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
