Getting your Trinity Audio player ready...
|
In a rapidly evolving world, security is paramount. With the pace of change accelerating, ensuring security solutions remain current and effective is more crucial than ever. Artificial Intelligence (AI) has emerged as a game-changer in security, and security camera systems are at the forefront of this transformation. AI is revolutionising security measures, enabling more efficient security solutions.
With this in mind, the National Institute of Standards and Technology (NIST) has evaluated software designed for Age Estimation and Verification (AEV) based on facial characteristics. This technology plays a crucial role in age-restricted activities, such as purchasing alcohol or accessing mature content online. By accurately estimating a person’s age, this technology helps enforce age restrictions, contributing to a safer and more secure environment for everyone.
Facial recognition technology is one of the most promising applications of AI for security cameras. This technology can accurately identify individuals flagged as suspicious or dangerous, significantly enhancing security protocols, especially in estimating individuals’ ages.
As security concerns continue to evolve, AI-driven technologies like facial recognition and AEV are poised to play an increasingly vital role in enhancing security measures across various sectors. It’s clear that staying ahead of the curve and embracing innovative security solutions are essential steps in safeguarding our communities and assets in an ever-changing world.
As digital technology improves, age estimation has become pivotal in age assurance programmes, which are increasingly incorporated into legislation and regulations worldwide. These programmes aim to restrict access to specific age groups for social media chat rooms and certain online and offline products, thus enhancing online safety for children.
The new NIST study, titled Face Analysis Technology Evaluation: Age Estimation and Verification (NIST IR 8525), assessed six algorithms submitted by developers. According to research, author Kayee Hanaoka said that the algorithms showed a wide range of performance levels, indicating significant room for improvement. “This is a partial snapshot of the age estimation field as it stood in late 2023,” Hanaoka noted, “but as AEV performance is closely tied to advancements in artificial intelligence, we expect the field to change rapidly.”
This study marks NIST’s first AEV evaluation in a decade, signalling a renewed, long-term commitment to regularly testing this technology. The last evaluation in 2014 was a one-time effort using a single database of about 6 million photos. The current study expands this to approximately 11.5 million photos from diverse U.S. government sources, reflecting various ages, genders, and regions of origin.
Key findings from the study include the absence of a single standout algorithm, with performance influenced by factors such as image quality, gender, and region of birth. Notably, AEV software has improved over the past decade, with mean absolute error rates dropping from 4.3 to 3.1 years on a shared database of visa photos. However, error rates remain higher for female faces than males, a trend observed in the 2014 evaluation.
The ongoing nature of NIST’s testing programme allows for continuous improvement and adaptation. The study authors are accepting new algorithm submissions and plan to release updates every four to six weeks. Future evaluations will explore additional questions, such as the potential for improved performance with prior photos of the same person. They will expand and diversify photo databases to cover applications like online safety better.
As artificial intelligence and digital technology continue to advance, NIST’s commitment to regularly evaluating AEV software will play a vital role in ensuring the security and accuracy of age-related digital technologies. This ongoing effort highlights the importance of adapting to new challenges and opportunities in the digital landscape, ultimately contributing to safer and more secure online environments.