25 January 2010 Fake fingerprint detection based on image analysis
Author Affiliations +
Fingerprint recognition systems have become prevalent in various security applications. However, recent studies have shown that it is not difficult to deceive the system with fake fingerprints made of silicon or gelatin. The fake fingerprints have almost the same ridge-valley patterns as ones of genuine fingerprints so that conventional systems are unable to detect fake fingerprints without a particular detection method. Many previous works against fake fingers required extra sensors; thus, they lacked practicality. This paper proposes a practical and effective method that detects fake fingerprints, using only an image sensor. Two criteria are introduced to differentiate genuine and fake fingerprints: the histogram distance and Fourier spectrum distance. In the proposed method, after identifying an input fingerprint of a user, the system computes two distances between the input and the reference that comes from the registered fingerprints of the user. Depending on the two distances, the system classifies the input as a genuine fingerprint or a fake. In the experiment, 2,400 fingerprint images including 1,600 fakes were tested, and the proposed method has shown a high recognition rate of 95%. The fake fingerprints were all accepted by a commercial system; thus, the use of these fake fingerprints qualifies the experiment.
© (2010) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Sang-il Jin, You-suk Bae, Hyun-ju Maeng, Hyun-suk Lee, "Fake fingerprint detection based on image analysis", Proc. SPIE 7536, Sensors, Cameras, and Systems for Industrial/Scientific Applications XI, 75360C (25 January 2010); doi: 10.1117/12.840216; https://doi.org/10.1117/12.840216

Back to Top