Constructing an AI Machine Vision Engine for Proprietary Identity Security With DICIO

Discover Synaptiq’s inventive use of AI and machine learning to encrypt the banking cycle for unbanked citizens worldwide, from enrollment to fraud detection.

Problem:

Passwords fail to protect personally identifiable and payment information time and again. DICIO wanted to use facial recognition and fingerprinting as dual forms of biometric authentication for individuals seeking to open a new bank account, many for the first time in their lives. DICIO needed facial recognition technology (“Crypto Facekey”) to determine both identities based on pre-existing databases as well as other aspects of well-being to help prevent fraud, spoofing, and other criminal activities. And all of this needed to happen in a secure and elegant app.

What can you learn from AI in banking?

Mathieu wanted best-in-class technology that his team would own and integrate into a platform for financial services companies. He sought out an experienced partner to build the necessary artificial intelligence (AI) machine vision engine and advise DICIO in developing the proprietary technology.

“Many people claim they have seen UFOs but it’s difficult to obtain unequivocal evidence. In the same way, it’s difficult to identify useful and successful implementations of emerging technologies, from AI to Blockchain to Image Processing. The real deals are built by unequivocally experienced technologists. The data scientists at Synaptiq provided the AI and Machine Vision experience we needed to build truly transformative tech that will get used by consumers, to the benefit of all.”

~ Stephane Mathieu, CEO, DICIO,

Solution:

Synaptiq helped DICIO CIO Stephane Mathieu and his team achieve their goal of developing a learning app with machine vision technology. To realize DICIO’s vision, Synaptiq explored sample images and videos compiled by the DICIO team. Based on this work, we proposed a few different ways to train and configure models for various tasks, including ways to measure stress based on determining heart rates from videos, an approach we uncovered in the process of working together. Over the course of our project, we fine-tuned and improved model accuracy. Finally, we delivered production-ready models and APIs and documentation on how to retrain models over time.

In summary, Synaptiq:

• Combined machine learning and machine vision technology into an app

• Proposed ways to train and configure models for various tasks

• Delivered production-ready models and APIs and documentation

Outcome:

Not only did we develop successful machine vision models and learning models for powerful and proprietary technology, but we also helped DICIO think through the operation of the product and where to plugin AI models to add value. We also advised on UX, APIs, overall product architecture.

Now when a resident of Mexico enrolls in a new bank account, s/he will have access to a device (smartphone or Kiosk) with a video camera and touch screen. Through a fine-tuned, easy-to-use series of screens, the individual will be asked to take a video of his/her face and a picture of his/her two fingers. The system will use these two inputs to ensure identity, eliminate the possibility of fraud and spoofing, and comply with regulatory and security requirements.

HUMANKIND OF IMPACT

100,000 residents of Mexico may now be able to bank for the very first time in their life.

100,000 residents of Mexico may now be able to bank for the very first time in their life.

HUMANKIND OF IMPACT