Mobile Verify®

Powering science

Artificial intelligence for maximum predictive power

A powerful combination of true identity document experts and sophisticated algorithms allows Mobile Verify to proof real-world identities in a smart and fully automated way.

Behind every transaction, our technology uses diverse AI analytics to achieve its maximum predictive power. From capturing an optimal image of the ID document to performing fully-automated authentication checks, or biometrically matching the ID document with a selfie, Mitek's science is built to return quick and conclusive results.

Learn more about Mobile Verify's AI-powered technology

Download whitepaper

Identity verified

Digital identity verification in less than 5 seconds

• Achieve faster approvals with full automation
• Establish trust in your digital platform
• Comply with AML/KYC regulations 

Document capture

It all starts with the quality of the image submitted

Mitek's capture SDK, MiSnap™, evaluates the capture conditions in order to interactively guide users through commands as they hover their smartphone over their ID document.

This real-time guidance is made possible by video frame analysis, which also allows for an auto-capture of the best possible image from the video stream. As a result, the first attempt is usually successful, which encourages widespread, rapid user adoption as well as a higher image acceptance rate.

Document classification

Once the image has been captured and submitted, Mobile Verify uses computer vision (CV) algorithms to match the image of the ID document to one of the thousands of formats of government-issued IDs from around the globe maintained in Mitek's cloud-based repository.

Data extraction

Through a combination of computer vision techniques, optical charater recognition (OCR) and rules, Mobile Verify compares the structure of the captured identity document against the corresponding document template in Mitek's repository.

Then, components such as printed bio-data, PDF417 barcode or the MRZ are accurately extracted and evaluated for consistency across the entire document.

Document authentication

In order to determine if the ID document is original and unaltered, Mitek's software unleashes hundreds of AI-based analytics on the image and extracted data, each one examining the ID in a different way to find anything suspicious.

Going beyond individual ML authentication checks, Mitek uses ensemble modeling for both authentication and decision-making. It is an intelligent approach exploiting the power of the learned data to provide accurate predictive results.

Biometric facial comparison

Once the ID document has been authenticated, Mobile Verify uses biometric facial comparison algorithms to automatically match the portrait extracted from the ID document with a selfie; proving that the person submitting the ID document is its rightful owner.

The software's ability to quickly and reliably make this match is substantiated by both "passive" and "active" anti-spoofing techniques. Passive tests operate in the background ensuring that the selfie is from a real person and not the result of a digital reproduction. Then the end-user is given a command to blink, as an active technique to attest a real-life interaction.

ID data analysis, fully-automated ML techniques, in-house ID expertise

While ML Algorithms learn from huge quantities of analyzed data, Mitek's ID document experts and data scientists also leverage their knowledge to investigate new trends and ways to improve the performance of Mobile Verify.

This process combines machines that quickly crunch through massive amounts of complex data with intel from our in-house ID document experts. It is a powerful engine to develop fully-automated forgery detection techniques.

Simultaneous AI-powered development & production

Because both the physical and digital worlds never stop changing, Mitek runs its development and production systems in parallel - giving our technology the ability to drive a rapid iteration cycle of continuous adaptation and improvement.