This scientific observation, a pillar of many criminal surveys and identification systems, has just been deeply revolutionized.
Artificial intelligence upsets our certainties on fingerprint identification. Indeed, a team from the University of Columbia discovered that all our fingers share a common biometric signature, invisible to the human eye but detectable by AI. This progress could breathe new life into criminal investigations and the biometric safety of our devices.
Principle of fingerprint identification:
Two fingerprints of different fingers of the same individual are different
Identification by fingerprints or dactyloscopy is a branch of biometrics, a science which makes it possible to identify people by means of their physical or biological characteristics.
Two individuals cannot have the same fingerprints, not even the real twins.
There are three large classes of forms of digital drawings: arcs, curls and vertics. The shape, size, number, number and arrangement of details, called minute, which characterize these patterns make each fingerprint unique.
We call “digital trace” or “latent papillary trace” a fingerprint raised on a crime scene. Its comparison with the fingerprints recorded in the police databases can make it possible to make connections between several offenses or to establish the presence of a suspect on a crime scene.
The revolution of artificial intelligence in the analysis of fingerprints:
The Columbia team has led to artificial intelligence over hundreds of fingerprints from different people. Over the course of this learning, AI has developed an amazing capacity to identify with precision exceeding 80 % that different fingerprints belonged to the same person.
This scientific breakthrough, published in the prestigious Revue Science Advances, reveals that AI is not limited to the minute traditionally used in typist.
Artificial intelligence has managed to observe what the human eye cannot detect, similarities in the corners and curvatures near the center of the fingerprints.
Revolutionary applications for justice and security:
This discovery could significantly improve criminal surveys. Connect an imprint found on a crime scene to a suspect whose imprint of another finger is only possessed. Thousands of unresolved cases could potentially be reopened.
The implications for Cold Cases are particularly promising.
How many fingerprints sealed in the premises of the police archives are waiting to be able to be linked to a suspect. This technology could establish links between different cases where different fingerprints, but belonging to the same person, have been found.
A new era in digital authentication:
The applications resulting from this discovery go far beyond the framework of judicial identification. Our electronic devices could soon recognize their owner, regardless of the finger used for biometric unlocking.
Indeed, this work can be useful in cases of digital authentication. Thanks to this methodology for the treatment of fingerprints, a person can register for the fingerprint scanner of their device with a finger and unlock it with any other
finger. This increases convenience, and it is also useful in cases where the original finger with which a person is registered becomes temporarily or definitively illegible, because it can always access its device with its other fingers.
The advantages of this daily user technology would be multiple:
• Unlocking devices or smartphones with any finger.
• Reinforced security against attempts at identity theft (scan of the 5 fingers of the hand at the same time).
• Simplicity of increased use on a daily basis.
• higher reliability in various environments.
• Access control in airports, businesses and sensitive buildings, with improved precision.
A new domain (biometrics) upset by AI:
This discovery questions a fundamental principle of typing established since the 19th century. If each digital imprint remains unique in its precise configuration for each finger, the existence of a signature common to all the fingers of an individual opens up fascinating perspectives.
The science of fingerprint identification enters a new era, where artificial intelligence reveals what the human eye could not perceive. This advance recalls how AI deeply transforms our most established scientific knowledge.
What man considered the only absolute truth
For more than a century is now nuanced by the higher analysis capacities of the algorithms of intelligenceartificial intelligence. A new page is written in the history of biometric identification.
The AI pushes the limits of our brain a little further.
In judicial matters, this discovery remains to be confirmed on a large sampling of individuals, and to be more perceptive (99.9%) in order to be recognized in court.
