Disruptive Computer Vision Provider Tops NIST Rankings with Lowest Error Rate in Critical Age Segments
ROC, the American-made computer vision and multimodal biometrics provider, emerged as a top performer in the recent NIST Face Analysis Technology Evaluation (FATE) for age estimation and verification, the first global benchmark of its kind in a decade.
In the study, which evaluates the ability of AI-powered tools to estimate a person’s age by looking at their face, ROC achieved the lowest Mean Absolute Error (MAE) of any vendor for individuals aged 18-24, with an impressive 2.3 year accuracy on mugshot data.
Mean absolute error (MAE) for mugshot subjects between the age 18 to 24 years old. Lower values are better.
Safeguarding the Next Generation
This success comes at a critical juncture, with skyrocketing demand for automated age verification due to new laws requiring increasingly stringent gatekeeping for age-restricted online content, services, and products. In the U.S., 18 states have passed age verification laws: Alabama, Arkansas, Florida, Georgia, Idaho, Indiana, Kansas, Louisiana, Mississippi, Montana, Nebraska, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Utah, and Virginia. At the same time, more and more question marks are being raised around the long-term risks of habitual social media and smartphone usage among our youth.
The overall results of NIST FATE, the most comprehensive and rigorous benchmark for face analysis technologies, placed ROC in the top tier of participants.
This validates ROC’s ability to estimate ages across a diverse audience set of human lifespans, geographic origins, and gender. The study included an expanded database of 11.5 million photos, including mugshots, border-crossing webcam images, and immigration headshots of people born in over 100 countries.
Mean absolute error (MAE) by gender for mugshot subjects between the age 18 to 30 years old. Lower values are better.
The Dawn of Digital Guardianship
As a research-first organization with deep scientific expertise, ROC leverages proprietary methods in computer vision and machine learning, powered by advancements in convolutional neural networks and other cutting-edge methods, to determine a person’s age with just their photo alone. The sophisticated models are trained on a massive, highly diverse dataset of responsibly sourced faces to minimize bias and produce the most precise age estimations available.
For ROC, the strong showing in NIST FATE demonstrates its innovative solutions for applications involving age-based face authentication, such as age-restricted access, age-appropriate content filtering, and age-sensitive marketing. ROC algorithms are highly accurate in estimating the ages of young adults, a challenging demographic due to rapid changes during adolescence. The algorithms excel at capturing the subtle differences in facial features that indicate age for individuals in their late teens and early twenties.
Automated Age Verification for Your Business
AI-driven age verification is used across industries for identity verification, marketing personalization, and the protection of minors.
- Precision and Protection: Ensure only legitimate customers above the legal age can access your age-restricted products and services. Reduce fraud and risk.
- Slash Manual Oversight and Costs: Eliminate the need for manual reviews, expedite customer journeys, and lower your cost of doing business.
- A Better Customer Experience: Streamline user onboarding with seamless and discrete real-time age verification.
- Ensure Full Legal Compliance: Meet all your adherence and compliance obligations by doing your utmost to prevent underage access to age-restricted offerings.
- Scale Faster: Optimize growth by scaling efficiently and serving more customers with the same resources.
ROC’s unique strength in qualifying young adults makes it an ideal partner for companies seeking the highest accuracy, data privacy, and ethical safeguards when verifying audiences – particularly those under 25.
Making the World Smarter, Safer, and Stronger
As the only American-made multimodal biometrics and computer vision provider, ROC is trusted by the U.S. military, law enforcement, and leading FinTech brands. The company has offices in Denver, CO, Morgantown, WV, and Grand Rapids, MI. ROC multimodal solutions deliver game-changing results with battle-tested reliability across NIST government testing, military applications, and 200M+ annual identity proofing transactions.