New Facial Recognition Algorithms Among Accuracy Leaders, According to NIST
Recent submissions of facial recognition algorithms to the U.S. National Institute of Standards and Technology (NIST) have placed several newcomers on the leaderboard for overall biometric accuracy. As of the latest update on July 5, algorithms from companies such as Megvii, Paravision, Kakao, Armatura, Verigram, Maxvision, and Omnigarde have secured their positions among the top 22 vendors.
Paravision, in particular, achieved impressive results, earning the second-most accurate score in the Visa-Kiosk category. Additionally, Paravision ranked within the top 5 in mugshot categories with databases containing 12 million and 1.6 million images. The company also excelled in matching webcam probe images to mugshots, as well as other categories such as Visa-Border and mugshots after 12-plus years. Moreover, Paravision ranked in the top 10 for border images compared after 10 years or more. Speaking about their success, Paravision CTO Charlie Rice commented on the wide-ranging usefulness of their Gen 6 technology.
NIST primarily uses the Visa-Border test as its primary sorting metric on the leaderboard. According to Paravision, there was a significant drop-off in performance after the top 5 vendors in this category. Paravision plans to deliver its Gen 6 facial recognition software to partners this summer, enabling them to deploy advanced and trusted solutions.
In addition to Paravision, other established developers such as Cloudwalk, SenseTime, NEC, and Idemia have also secured top positions on the leaderboard, as expected. However, Omnigarde, an up-and-coming developer, also celebrated its achievement as one of the leading U.S. companies in the biometric technology sector.
NIST’s July 5 Face Recognition Vendor Test report evaluated more than 350 algorithms, showcasing the latest advancements in facial recognition technology. The report provides valuable insights into the accuracy and capabilities of these algorithms, aiding organizations in selecting the most suitable solutions for their needs.
As the field of facial recognition continues to evolve, these latest algorithm submissions highlight the growing competition and the emergence of new players in the industry. Improved accuracy and reliability in biometric technologies contribute to enhanced security measures and the development of innovative applications. It will be fascinating to see how these advancements impact various industries worldwide.
Overall, NIST’s evaluation and benchmarking efforts play a crucial role in ensuring the integrity and performance of facial recognition technology. With the continued advancement of algorithms, the industry is poised for further growth and innovation, fueling the adoption of facial recognition in numerous sectors, including law enforcement, border control, and identity verification.
The results of the NIST Face Recognition Vendor Test serve as a testament to the ongoing technological advancements and the potential of facial recognition to revolutionize security and identity management practices. It is an exciting time for the facial recognition industry as it continues to push boundaries and deliver more accurate and reliable solutions to address a wide range of challenges and opportunities.
