IDnow just won a new research project, funded by the Brittany region and by the Images et Réseaux cluster. This project will last 10 months and begin in March 2021, in collaboration with the b<>com institute. It aims to develop a new fraud detection method based on the fingerprints left by the users.
A better fraud detection
First, the document is automatically analyzed to extract its information (photo, name, birth date…) and check its authenticity. Then, the user from the video is compared to the photo on the document to verify it is the same person.
Today, this service keeps no record of verification history. However, this would allow for more effective fraud detection, as many fraudsters reoffend and reuse identity elements from one attempt to the next. It would also make it possible to detect the use of the same device to check many different identities.
FIFRAUD aims to tackle this issue and proposes to collect the fingerprint associated with each ID verification request. It will be compared with the fingerprints left by previous inquiries, to be able to link several suspicious requests.
In addition to the technical developments to calculate and compare fingerprints, led by the IDnow Research and Development team, several challenges have to be met:
- The service must be GDPR compliant, which means that the data have to be anonymized;
- Current processing times, inferior to 10s per request, needs to be maintained;
- The system will have to be able to handle large fluctuations in the number of requests, especially during peak demand periods.
Collaboration with b<>com Technology Research Institute
IDnow, FIFRAUD project carrier, will produce the technical bricks with the support of b<>com Technology Research Institute. ARIADNEXT will be in charge of the modules to detect the repetition of the same identity elements. b<>com will supply and adapt its terminal authentication solution, in order to be able to collect the fingerprint of a user on a web browser or on a cell phone.
The developed solutions will be integrated into IDCheck.io.
Interested in collaborating with IDnow?
Research and Innovation is at the heart of IDnow’s activity since the creation of the company. Our teams of experts (image analysis, character recognition, biometry, Artificial Intelligence, programming…) want to push the limits of ID verification and fraud detection to propose even more efficient technologies. We are convinced that sharing knowledge brings us further, that is why we welcome any collaboration offer. Feel free to contact us!
FIFRAUD is supported by Brittany Region.