Deep scanning technology for next-level fingerprint sensors
Ever since they started appearing on smartphones in 2007, fingerprint sensors have been proving quite popular. However, they are far from perfect: any change or damage to the finger will prevent current scanners from recognising it. ‘In the case of altered fingerprints, existing sensors will not be able to capture a fingerprint image of sufficiently high quality for automatic processing in order to perform recognition,’ explains Stéphane Revelin, Programme Manager at Idemia and coordinator of the INGRESS (Innovative Technology for Fingerprint Live Scanners) project. Most smartphone owners who have tried scanning their fingerprints after they have been in contact with liquid, or after an intensive day of manual work, will probably be familiar with this scenario. The two technologies developed under INGRESS were created with these problems in mind. Using ultrasound and Optical coherence tomography (OCT), respectively, they are capable of capturing internal fingerprints – that is, fingerprints under the surface of the skin. ‘Some studies have showed that the fingerprint pattern is present in the epidermis. The goal of our innovative sensor is to capture this pattern and get a fingerprint image that is usable for fingerprint recognition with legacy systems and algorithms,’ Revelin says. ‘Besides, a third technology has been developed under INGRESS: a slim sensor based on Printed organic electronics (POE). It has several advantages compared to current sensors: a lower cost and a very small form factor even for large surfaces.’ Developing such technologies was not easy. Over the course of the project, the team first found out that OCT technology was better at recognising internal fingerprints than ultrasound. The technology could capture sub-surface fingerprint images with a high enough quality to allow for processing. From there on, the team could demonstrate that, in the case of a damaged fingerprint, recognition was not affected. ‘We have made tremendous progress,’ Revelin enthuses. ‘However, we cannot yet pretend to outperform current solutions. There is still some work required to improve the quality and the accuracy of our system. Our benchmarks indicate that the legacy sensors are still providing more accurate results, but OCT is quite close.’ The team goal for POE was to achieve a resolution of 500 dpi for the sensor, compared to state-of-the-art scanners which provide 64 dpi images. Whilst this objective couldn’t be reached in the timeframe set for the project, the team came very close to developing a prototype. Just like for OCT, they intend to pursue their work in order to get closer to the future industrial phase. When they finally reach this level of maturity, INGRESS technologies may find use notably in border control, as Revelin explains. ‘Border police are facing a number of situations where voluntarily or involuntarily damaged fingerprints make it very difficult to access information about the traveller. Sub-surface sensors might come in very handy in these situations. When mature, POE technology could even be integrated into the screen of a smartphone or tablet.’ If successful here, POE will certainly find a prominent place in the highly competitive smartphone industry.
Keywords
INGRES, fingerprint, POE, smartphone, security, scanner, fingerprint alteration, sensor, internal fingerprint, OCT