Limits...
Fingerprint Liveness Detection in the Presence of Capable Intruders.

Sequeira AF, Cardoso JS - Sensors (Basel) (2015)

Bottom Line: We approach the design by modeling the distribution of the live samples and predicting as fake the samples very unlikely according to that model.Our experiments compare the performance of the supervised approaches with the semi-supervised ones that rely solely on the live samples.The results obtained differ from the ones obtained by the more standard approaches which reinforces our conviction that the results in the literature are misleadingly estimating the true vulnerability of the biometric system.

View Article: PubMed Central - PubMed

Affiliation: INESC TEC-INESC Technology and Science, Campus da FEUP, Rua Dr. Roberto Frias, Porto 4200-465, Portugal. afps@inesctec.pt.

ABSTRACT
Fingerprint liveness detection methods have been developed as an attempt to overcome the vulnerability of fingerprint biometric systems to spoofing attacks. Traditional approaches have been quite optimistic about the behavior of the intruder assuming the use of a previously known material. This assumption has led to the use of supervised techniques to estimate the performance of the methods, using both live and spoof samples to train the predictive models and evaluate each type of fake samples individually. Additionally, the background was often included in the sample representation, completely distorting the decision process. Therefore, we propose that an automatic segmentation step should be performed to isolate the fingerprint from the background and truly decide on the liveness of the fingerprint and not on the characteristics of the background. Also, we argue that one cannot aim to model the fake samples completely since the material used by the intruder is unknown beforehand. We approach the design by modeling the distribution of the live samples and predicting as fake the samples very unlikely according to that model. Our experiments compare the performance of the supervised approaches with the semi-supervised ones that rely solely on the live samples. The results obtained differ from the ones obtained by the more standard approaches which reinforces our conviction that the results in the literature are misleadingly estimating the true vulnerability of the biometric system.

No MeSH data available.


Finger Play-Doh mold and silicon model.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4507655&req=5

f2-sensors-15-14615: Finger Play-Doh mold and silicon model.

Mentions: Biometric recognition systems in general, and FRS in particular, can be spoofed by presenting fake or altered samples of the biometric trait to the sensor used in a specific system. Liveness detection techniques and alteration detection methods are both methods included in the presentation attack detection (PAD) methods [5]. Concerning spoofing attacks with fake samples, the fake samples can be acquired with or without user cooperation: an authorized user may help an hacker to create a clone of his fingerprint; or the fingerprint may be obtained from a glass or other surface (latent fingerprints) [2,6]. Latent fingerprints can be painted with a dye or powder and then “lifted” with tape or glue. However, these prints have, usually, low quality as they can be incomplete or smudged and thus are not very accurate. The easiest way of creating a fake sample is by printing the fingerprint image into a transparent paper. A more successful method is to create a 3D fake model with the fingerprint stamped on it. This can be done by creating a mold that is then filled with a substance (silicon, gelatin, Play-Doh, wax, glue, plastic). Then this mold is used to create a thick or thin mold that an intruder can use, as depicted in Figure 2.


Fingerprint Liveness Detection in the Presence of Capable Intruders.

Sequeira AF, Cardoso JS - Sensors (Basel) (2015)

Finger Play-Doh mold and silicon model.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4507655&req=5

f2-sensors-15-14615: Finger Play-Doh mold and silicon model.
Mentions: Biometric recognition systems in general, and FRS in particular, can be spoofed by presenting fake or altered samples of the biometric trait to the sensor used in a specific system. Liveness detection techniques and alteration detection methods are both methods included in the presentation attack detection (PAD) methods [5]. Concerning spoofing attacks with fake samples, the fake samples can be acquired with or without user cooperation: an authorized user may help an hacker to create a clone of his fingerprint; or the fingerprint may be obtained from a glass or other surface (latent fingerprints) [2,6]. Latent fingerprints can be painted with a dye or powder and then “lifted” with tape or glue. However, these prints have, usually, low quality as they can be incomplete or smudged and thus are not very accurate. The easiest way of creating a fake sample is by printing the fingerprint image into a transparent paper. A more successful method is to create a 3D fake model with the fingerprint stamped on it. This can be done by creating a mold that is then filled with a substance (silicon, gelatin, Play-Doh, wax, glue, plastic). Then this mold is used to create a thick or thin mold that an intruder can use, as depicted in Figure 2.

Bottom Line: We approach the design by modeling the distribution of the live samples and predicting as fake the samples very unlikely according to that model.Our experiments compare the performance of the supervised approaches with the semi-supervised ones that rely solely on the live samples.The results obtained differ from the ones obtained by the more standard approaches which reinforces our conviction that the results in the literature are misleadingly estimating the true vulnerability of the biometric system.

View Article: PubMed Central - PubMed

Affiliation: INESC TEC-INESC Technology and Science, Campus da FEUP, Rua Dr. Roberto Frias, Porto 4200-465, Portugal. afps@inesctec.pt.

ABSTRACT
Fingerprint liveness detection methods have been developed as an attempt to overcome the vulnerability of fingerprint biometric systems to spoofing attacks. Traditional approaches have been quite optimistic about the behavior of the intruder assuming the use of a previously known material. This assumption has led to the use of supervised techniques to estimate the performance of the methods, using both live and spoof samples to train the predictive models and evaluate each type of fake samples individually. Additionally, the background was often included in the sample representation, completely distorting the decision process. Therefore, we propose that an automatic segmentation step should be performed to isolate the fingerprint from the background and truly decide on the liveness of the fingerprint and not on the characteristics of the background. Also, we argue that one cannot aim to model the fake samples completely since the material used by the intruder is unknown beforehand. We approach the design by modeling the distribution of the live samples and predicting as fake the samples very unlikely according to that model. Our experiments compare the performance of the supervised approaches with the semi-supervised ones that rely solely on the live samples. The results obtained differ from the ones obtained by the more standard approaches which reinforces our conviction that the results in the literature are misleadingly estimating the true vulnerability of the biometric system.

No MeSH data available.