Smart Engines Launches AI Scanner for Handwritten Credit Card Numbers

Smart Engines has launched an AI solution for reading credit card numbers recorded in any format. Banking customers simply point their smartphone camera at a printed or handwritten credit card number to make a transfer or pay for a purchase at a point of sale. The neural network recognizes the number combination in real-time and automatically enters it into the payment form. The user only has to confirm the transaction.

The detector captures the credit card number in a photo, scan or in real-time, even when other text is printed or written around it, and then reads it. The neural network recognizes 16-digit credit card numbers, which are the most common today, as well as cards with 13, 15, 18 or even 19 digits.

This unique and unprecedented AI solution has been developed by Smart Engines. The functionality has become part of the Smart Code Engine 2.0 product. The solution is implemented in mobile applications distributed in the App Store and Google Play. The software is also installed in banks’ PWA applications using WebAssembly technology. Smart Engines’ scanning software solutions are based on GreenOCR® technology with environmentally friendly algorithms that do not require an Internet connection, the presence of a GPU or powerful computers on the customer side.

At the time of this announcement, Vladimir Arlazarov, CEO of Smart Engines, PhD in Computer Science shared more insights on the role of Artificial Intelligence-based scanner for credit card numbers. Here’s a snapshot of our conversation with the CEO of Smart Engine.

AiThority: What are the advantages of using an AI-based scanner for handwritten credit card numbers? 

Vladimir Arlazarov [VA]: “Credit card numbers are often used to facilitate person-to-person or person-to-store payments, and while some solutions existed before which allowed to quickly fill out the recipient’s bank card number in a mobile banking application, for that to work the recipient have to present the credit or debit card itself. AI-based scanning of printed or handwritten credit or debit card numbers allows to perform transfers by simply pointing the camera to a written number on a website, in an e-mail, or on a piece of paper in a local small business.”

AiThority: What is the data science involved in the development of your AI product?

[VA]: “To develop such AI product we had to develop computer vision modules dedicated to printed and handwritten digits detection in an arbitrary scene, and precise recognition of the digits in a format of a bank card, bearing in mind a scarce data availability. Thus, the core data science methods we use in this product, as in our other products, are related to synthetic training data generation, data augmentation, and, of course, machine learning for the computer vision algorithms themselves.”

AiThority: Is the use of AI safe for the financial and banking services industry? What are the critical differentiators in this industry that make AI so useful for customer and business owners?

[VA]: “The safety of AI products is conditioned on the approach used by their developers. I would say that the crucial differentiator in this industry is the ethics of data handling – whether the handling of the data in an effort to build a useful product compromises the security and privacy of the data owner or not. 

Taking bank cards scanning solution as an example – in this product, as in all our products, the analyzed video frames does not leave the device where the recognition takes place, thus the possibility that the scanned debit card numbers would get into wrong hands is eliminated. With such an approach the usage of this product is as safe as a card number entry box in your bank’s application. The usage of AI products in the banking industry is safe, if the developer and the client understand who and how they are helping, and what added risks they introduce. You wouldn’t ask if the usage of a keyboard is safe for online banking website – because this is just your data entry tool, and you can trust the bank not to send the data you type in to a third party. The same here – if the AI product performs the recognition on the end-user device itself, without sending the images or video frames anywhere, then it is as safe as a keyboard.”

Making Credit Card Payments Convenient for Users

Smart Engines’ credit card number scanner will significantly improve the convenience of data entry when making payments through any banking application and online banks in the browser. The handwritten credit card number recognition process takes place automatically, in real-time and directly on the user’s device. Customer data is not transmitted to third-party services, making the entry of payment information secure for both the user and the mobile application owner.

“The credit card number scanner is another cutting-edge technology from the Smart Engines team of scientists. It is the highlight of the updated version of our Smart Code Engine 2.0 product. The credit card number scanner completely changes the input process and makes it more convenient than ever. Nobody likes to type in a number by hand. Smart Engines OCR takes the hassle out of the process by instantly recognizing the number combination and entering it into the payment form. All they have to do is confirm the payment,” commented Vladimir Arlazarov, CEO of Smart Engines, PhD in Computer Science.

Smart Engines is a leading developer of systems that recognize and input data from ID cards, credit cards, OR codes and barcodes presented to the camera in real time, as well as from photos and scans. Smart Engines develops its own software products – Smart ID Engine, Smart Code Engine, Smart Document Engine, which solve all business tasks in the field of document management. In addition, Smart Engines has developed its own tomographic reconstruction system, Smart Tomo Engine.

[To share your insights with us, please write to sghosh@martechseries.com]