printer friendly version

Live fingerprint detection

An introduction to Suprema’s live fingerprint detection (LFD) technology.

Suprema’s LFD technology is based on the fact that dynamic and static image characteristics of fake fingers can be distinguished from those of live fingers. By the advanced analysis algorithm to catch the abnormalities in dynamic changing pattern of fingerprints images, and several static features showing liveness or unnaturalness of fingers, fake fingers are clearly distinguished from those of live fingers.

This new LFD technology provides a cheap and effective solution to protect the fingerprint system from attack via fake fingerprints. The LFD function is already applied to our brand-new USB fingerprint scanner, Biomini-Plus and is planned to be applied to our embedded module product, SFM series, in the near future.

Dynamic changing pattern analysis

When fingers are gradually contacting on the sensor surface, live fingers make natural changing patterns of area, intensity, and movement, but fake fingers make unnatural changing patterns of separated areas, partially dark area, distorted boundary shape, and large movement of core part. By catching these abnormalities in dynamic changing patterns from continuous fingerprint images, fake fingers are distinguished from live fingers. Specifically, this method is very powerful in rejecting out the fake fingers made by hard materials, eg, paper, film, clay, and hard rubber.

Liveness feature analysis

In fingerprint images, there are several localised features which reveal the liveness of fingers, pore distribution, ridge sharpness, regularity of ridge-valley boundary etc. These localised liveness features are normally too minute and elaborate to be copied by simple and soft faking materials, eg, silicon, rubber, and gelatine. Our high performing imaging sensor can capture high quality fingerprint images, and various local liveness features are precisely described by our advanced analysis algorithm.

Unnaturalness feature analysis

Usually, it is very hard to make perfect fake fingers and almost every fake finger cannot avoid revealing its unnaturalness – unnaturally sharp boundaries, too many white blobs or too large black blobs within fingerprint area, abnormal peaks in histogram distribution etc. By observing the mixture of numerous unnaturalness features, naïve elementary fake fingers are rejected perfectly.

Share this article:

Further reading: