ROC.ai Sets New Iris Recognition Benchmark for Combined Speed and Accuracy, Launching Globally Competitive Algorithm in Record Time
In the prestigious NIST IREX rankings, ROC.ai’s algorithm proved to be 100x faster than NEC, IDEMIA, or Thales – with an accuracy in searching two eyes (left and right iris) of 99.3% at a false positive ID rate of 1.0%.
Continuing their blockbuster drive to put American-made computer vision algorithms at the top of global performance, Rank One Computing (ROC.ai) has announced the launch of their groundbreaking iris recognition solution. By leveraging their cutting-edge AI, ML, and deep learning expertise, ROC.ai was able to develop an iris matching algorithm that outperforms global competitors in just a few short months.
“While not surprised, I am remarkably impressed with our team’s ability to build such a powerful set of algorithms in such a short time. We announced our first fingerprint algorithms earlier this year and now have one of the leading iris algorithms. We’re developing from the ground up, here in the U.S., and competing head-to-head with established foreign biometric giants.”
Scott Swann, CEO of ROC.ai
Iris recognition, a form of biometric authentication that relies on the unique patterns in a person’s iris, has been gaining traction for its high accuracy and security. It has critical applications across multiple sectors, from border control and national security to access control and financial services. However, achieving a high level of accuracy is no easy feat, as it requires advanced algorithms that can handle the complex and delicate task of matching iris patterns. The success of ROC.ai in this field is a testament to their prowess in artificial intelligence, deep learning, and the development of complex neural networks that replicate the human visual system.
“The ROC team’s journey began less than six months ago with a commitment to revolutionize the field of iris recognition. We embarked on a mission to develop a U.S.-native algorithm that would establish a new standard for accuracy and speed. We are delighted to have delivered upon this goal in our first submission and in and record time frame.”
Blake Moore, ROC.ai VP of Product and Customer Success.
NIST IREX 10: Computation Time and Quality Assessment
Iris recognition search speed and error rate comparison between ROC.ai and other NIST IREX participants.
When tested on a 1M iris database, ROC.ai achieved a mean search speed of 0.11s. In contrast, no other vendor achieved a search speed of less than 1 second, with the majority falling above the 10-second mark. This demonstrates ROC.ai’s ability to outperform other vendors by 100x in iris database searches, achieving exceptional speed and highly competitive accuracy. Source: https://pages.nist.gov/IREX10
A Game-Changing Approach to Iris Recognition
The National Institute of Standards and Technology (NIST) conducts its IREX tests to evaluate the quality and reliability performance of iris recognition algorithms worldwide. These rankings are highly regarded in the biometric community and signify an algorithm’s quality and reliability.
By securing a NIST ranking in an unprecedented timeframe, ROC.ai demonstrated the efficiency and robustness of their highly innovative, scalable approach.
Traditionally, it takes years for biometric companies to develop algorithms that can compete on a global scale. What is most noteworthy about ROC.ai’s achievement is that they managed to do it in just a few months. This accelerated pace not only underscores the technical expertise of the ROC team, but also reflects the incredible potential of American-made technology in transforming the global biometrics industry.
“Our approach to developing iris recognition capabilities was quite similar to past efforts that have delivered the industry-leading face and fingerprint recognition algorithms that ROC.ai is renowned for. It is a combination of the pattern recognition scientific principles espoused by the seminal textbook Pattern Classification, a blue-collar work ethic and mentality, and software engineering trade secrets that for nearly a decade now have landed ROC as the fastest and most efficient computer vision algorithm developer in the world.”
Brendan Klare, ROC.ai Chief Scientist and Co-Founder
A New Dawn for American-Made Biometrics
The groundbreaking achievement of ROC.ai in developing a NIST-ranked iris recognition algorithm in record time is testament to the game changing power of American innovation. Slowly but surely, American companies are closing the gap in global biometrics capabilities by incorporating novel, innovative approaches, new standards of ethical and technical excellence, and cutting-edge AI, ML, and deep learning expertise. With their visionary approach, unwavering dedication to data privacy, and an industry-leading code of ethics, ROC.ai is showing the world that the future of biometric security is brighter than ever.
A World Leader in Computer Vision and Biometrics
At ROC.ai, we believe in the power of technology to make the world smarter, safer, and stronger – without bias or threats to privacy. Trusted by the U.S. military, law enforcement, and FinTech, our multimodal solutions deliver game-changing results with battle-tested reliability across NIST government testing, military applications, and 200M+ annual identity proofing transactions. ROC.ai is employee-owned, ethics-driven, and 100% made in America. Our company has offices in Denver, Colorado and Morgantown, West Virginia.