Post Bank has implemented VisionLabs protection against biometric fraud using photos

2022: Implementation of biometric fraud protection

Post Bank implemented protection against biometric fraud using ready-made photos and images from mobile device screens. VisionLabs announced this on September 13, 2022. VisionLabs' Liveness algorithms check whether a person is actually in front of the camera, and not their image, thereby protecting against internal and external fraud. The solution is scaled to more than 26 thousand customer service points in the branches of the Russian Post.

The updated functionality is implemented on the basis of the LUNA PASS service from VisionLabs, which allows you to take photos through a web application. When photographing customers for service or registration of applications, the system checks the quality of the received photo by more than 15 parameters, including the absence of illumination, sufficient illumination, head rotation, etc., and also controls that the process takes place without replacing the image. This allows you to protect yourself from scammers who are trying to take out a loan or withdraw cash from someone else's photo.

Pochta Bank has been using its own internal biometric identification of customers by face since its inception. It is designed to protect against fraudulent activities and illegal access to the account both in service offices and in the mobile application. A sample image of each client's face is recorded at the first visit to the bank, and later this impression is used to identify it as an additional factor along with presenting a passport. A bank employee can gain access to a client's account and conduct all transactions only with the personal presence of the client and after his identification by photographing, thus unauthorized access to accounts is completely excluded. Also in the mobile application, customers can independently confirm atypical operations using selfies, which significantly increases the security of using the application.

Our biometric system has been reliably protecting customer funds both in all service points and in the application for many years. In 2021 alone, thanks to its use, we identified and prevented the execution of more than 10 thousand loan applications with suspected fraud in the amount of more than 4 billion rubles. The introduction of "smart" liveness control technology has made it even more reliable and effective, "said Svyatoslav Emelyanov, Deputy President and Chairman of the Board of Post Bank.

Liveness algorithms are also used to authorize employees in the bank's CRM system and protect against internal fraud. When you log into a face biometrics account, LUNA PASS verifies the image's authenticity to avoid unauthorized access to internal information and customer accounts. If the Liveness check fails, access to the CRM system will be denied.

Protection against forgery through printed photos, images or videos from mobile device screens is made possible by VisionLabs computer vision technologies. Neural networks analyze the obtained image for background inconsistencies, presence of glare, face image distortions, and perspective distortions. At the same time, only one frame is enough - a special technology does not require additional interaction or long video, which makes checking fast and convenient. The process takes seconds and is invisible to the user.

The introduction of protection against biometric spoofing attacks in one frame allows you to increase the security of financial transactions. Computer vision technologies as part of LUNA PASS are completely our own development, "said Dmitry Markov, CEO of VisionLabs.

In addition, LUNA PASS does not require significant hardware power and minimizes link requirements. The VisionLabs team has done significant work to optimize and configure the internal services of the product so that it works even at low data rates and high delays in the data channel.