Fingerprint recognition cracked? Deepening subcutaneous to improve lock safety

2024-03-13

Applying each person's unique fingerprint as a feature of identity recognition to various "locks" is no longer just a plot in science fiction works. Whether it is fingerprint lock or fingerprint unlocking of electronic products such as mobile phones, fingerprint recognition has already been widely used in daily life. This type of "key" that is difficult for others to replicate greatly enhances the security of modern electronic locks.

However, since ancient times, lock and unlock technologies have been constantly competing with each other. As fingerprints become special keys, the cracking technology of fingerprint locks is also constantly developing. In film and television works, it is not uncommon to use special materials to imitate fingerprints and crack fingerprint locks. However, in real life, with the development of technologies such as 3D printing, the method of imitating fingerprints has also become a reality. When the security of fingerprint locks is challenged, as a widely used biometric technology, fingerprint recognition technology also needs further development.

How to enable fingerprint recognition technology to distinguish counterfeit fingerprints? The research team from the School of Computer Science and Technology at Zhejiang University of Technology has collaborated with the First Research Institute of the Ministry of Public Security to propose a solution - to deepen the recognition depth of fingerprint recognition to 1 to 3 millimeters below the surface of the finger skin. In addition to fingerprint information on the finger surface, it can also collect high-resolution three-dimensional subcutaneous structure information.

Fingerprints are the protrusions of the dermal tissue between the epidermis and dermis towards the epidermis. The fingerprints on the surface of the skin are only the contours of the top stratum corneum of the epidermis, while internal fingerprints still exist below the stratum corneum. The internal fingerprint of the fingertip is the source of external fingerprints. On the one hand, it is protected by the stratum corneum, which is not easily damaged, and is not affected by external environmental factors such as dust and sweat. Collecting internal fingerprints avoids the problems of existing fingerprint recognition technologies obtaining surface patterns and poor recognition accuracy. On the other hand, current fingerprint counterfeiting technology can only forge surface fingerprint patterns, without internal fingerprints below the stratum corneum and internal tissue structure information such as sweat glands, which can greatly improve the security of existing fingerprint recognition technology.

The new generation of multimodal biological feature acquisition technology developed by Zhejiang University of Technology and the First Research Institute of the Ministry of Public Security can stereoscopically scan images of internal and external fingerprints and sweat gland sweat pores. This not only strengthens the defense against fingerprint imitation technology, but also enhances the effectiveness of fingerprint information collection, improving recognition accuracy. In the experiment, the device can identify specially made fake fingers and fingerprint films without internal fingerprint and subcutaneous three-dimensional structural data. If the fingerprint film is extremely thin, it can even extract real fingerprint information.

It is reported that this new fingerprint recognition device includes prisms, filters, fluorescent light sources, industrial cameras, fiber optic couplers, and other devices. It uses total internal reflection (TIR) technology to collect external fingerprints, while synchronously collecting internal fingerprints using optical coherence tomography (OCT). Two completely different optical systems are fused through a self-made trapezoidal prism to achieve the same measurement area and synchronous acquisition. To ensure the area of optical coherence tomography imaging, the team also used small spot fiber collimators and telephoto lenses. In addition, the research team has specifically developed deep adaptive multi convolutional fusion technology to address the issue of varying thickness of the stratum corneum for each individual. Even if the depth of each individual's human tissue is different, intelligent algorithms can accurately obtain subcutaneous tissue structure information.

At present, most domestic smartphones adopt fingerprint unlocking technology, and the market size of fingerprint smart door locks is also constantly expanding. With the increasing application of fingerprint recognition technology, its security and information collection capabilities also need to be given more attention. Not only do we need to enhance our ability to identify fake fingerprints, but we also need to ensure that elderly people, children, or those born with indistinct fingerprints can unlock them smoothly. The new technology provides the possibility to achieve these two capabilities. Currently, this technology has not been put into practice, but the team's next step is to collaborate with enterprises to continuously verify device performance. Perhaps soon, we will be able to use a more secure fingerprint lock.

Japanese scientists use 3D printing technology to create food with marble pattern effects