TOKYO -- A team of researchers in Japan has developed fingerprint recognition technology that detects sweat pores in fingers to increase the reliability of authentication.
Researchers at the University of Tokyo and DDS, a biometric technology startup, believe the breakthrough improves precision more than tenfold compared to current identification methods.
Fingerprint identification is widely used in smartphones, but experts believe the current technology lacks precision, leaving phones vulnerable to forged or misidentified fingerprints.
"Existing fingerprint authentication technology is limited in terms of precision," said University of Tokyo professor Taizo Umezaki, explaining the reason behind development of the new technology. "Smartphones are now used for a wide range of applications, like electronic payments, so we need a simpler and more precise way of identification."
Current identification works primarily by recognizing patterns, such as swirls, in fingerprints. The process is enhanced by recognition of the forking and end points of lines.
The small scanners used in smartphones are limited in the amount of detail they can read, increasing the risk of misidentification or forgery. The probability of misidentification for larger scanners found, for example, at room entrances is about one in a million. For smartphones, this worsens to around one in 50,000.
A research team at Michigan State University in the U.S. said it was able to unlock smartphones using images of fingerprints printed on an inkjet printer. Phones have also been unlocked by people other than the legitimate users.
To improve precision, the new technology adds recognition of sweat pores to the identification process. Sweat pores, like the fingerprint itself, differ from person to person. The new process measures pore positions relative to each other and to the lines in a fingerprint. Owing to the large number of sweat pores in a finger, even small scanners will be able to collect more information.
A major technical hurdle was actually recognizing the tiny pores, as capacitive sensors -- the type commonly used today -- have limited resolution. To overcome this, DDS developed a scanner that combines a thin glass plate, image sensor and light-emitting diode to illuminate the fingerprint, providing the ability to discern minuscule features.
The scanner is 6.6mm long, 4.8mm wide and about 0.6mm thick, with a resolution up to 3,000 pixels per inch -- a significant improvement over the roughly 500 ppi of existing scanners. Umezaki and DDS also developed the scanner's software.
One issue yet to be addressed is the high cost of the device, which at present precludes its use in smartphones. But the team believes the cost will decrease when the scanners are mass-produced. Scanners in smartphones today are priced between 100 and 1,000 yen (93 cents to $9.3) each. Plans are in the works to introduce the new device to smartphone manufacturers.
The team also hopes the technology can be adapted to work on display panels, allowing it to be installed in the growing number of smartphones that are eliminating the home button. They also think it can be used in other applications, such as car and home locks.
Many types of biometric identification -- vein, facial, iris and voice recognition -- are available. Apple uses 3D facial recognition in its iPhone X, while Samsung Electronics employs a combination of facial and iris recognition in its Galaxy S9. These methods may eventually replace fingerprint identification, but doubts remain about their reliability.
For example, technology that relies on a smartphone's camera, such as facial recognition, requires that the camera have adequate optical performance. Sometimes smartphones fail to identify the proper face, or cannot tell the difference between twins.
Vein recognition is believed to be highly accurate, but is limited to machines such as ATMs due to the size of the scanner.
Analysts say fingerprint recognition will remain a key technology for different areas, giving the new sensor a good chance of catching on.