AI Computer Vision
Protect access to your workplace with facial security
3D body (skeletal) tracking middleware
Edge AI hardware
for face and body tracking
Facial recognition platform for video processing
Advanced face liveness detection for digital onboarding and KYC
Scalable fasial recognition API for on-pream or AWS cloud deployment
Facial and body recognition library for server, mobile and embedded solutions
3DiVi Inc., founded in 2011, is one of the leading developers of AI and machine learning (ML) technologies for computer vision.
3DiVi News

Our Quality Assessment Algorithm Recognized by NIST FATE Quality Ratings

At 3DiVi, we continuously strive to push the boundaries of technology, and our recent advancements in the quality assessment algorithm for face recognition are a testament to this commitment. Our cutting-edge algorithm, integrated into our QAA (Quality Assessment Algorithm) product, has shown remarkable performance improvements, validated by the NIST FATE (Face Recognition Vendor Test) Quality ratings.

Here’s a detailed look at our latest achievements:

  • Comprehensive Quality Metrics Analysis: our enhanced algorithm now evaluates 13 out of 22 possible quality metrics, providing a thorough and nuanced analysis of facial recognition quality. This comprehensive assessment allows us to deliver more reliable and precise results, ensuring higher accuracy and consistency in various applications.

  • Optimized Processing Time: one of the significant improvements is the processing speed of our algorithm. The time to analyze a single image is fraction of second, placing us on par with the industry standard. This optimization ensures that our clients receive quick and efficient quality assessments without compromising on accuracy.

  • Top-tier Performance in Detection Metrics: in the realm of face detection, our algorithm has achieved outstanding rankings:
— missed detections: we are ranked in the top 3 globally. This high ranking indicates our algorithm's exceptional capability in accurately identifying faces, minimizing missed detections and thereby increasing reliability.

  • Superior Angle Estimation: accurate determination of facial angles is crucial for numerous applications. Our algorithm excels in estimating Yaw, Roll, and Pitch angles, securing a position in the top 10 worldwide. This capability enhances the accuracy of face recognition systems, particularly in environments where facial orientation can vary significantly.

  • Leading in Background Uniformity Assessment: uniformity of the background in facial images is a critical quality metric, and our algorithm is among the top 3 in the world for this criterion. This achievement highlights our proficiency in evaluating and ensuring the quality of background conditions, which is essential for accurate and consistent face recognition performance.

  • Overall Performance Excellence: in the comprehensive evaluation by NIST FATE, considering various metrics and discarding the bottom 5% of the worst cases, our algorithm ranks in the top 5 globally. This overall ranking reflects our commitment to delivering top-tier performance and reliability across multiple quality metrics.

Our continuous innovation and dedication to excellence have enabled us to make significant strides in the quality assessment of facial recognition systems. The recent improvements in our QAA product, validated by the NIST FATE Quality ratings, demonstrate our leadership and commitment to providing state-of-the-art solutions to our clients. As we move forward, we remain focused on further enhancing our technologies to meet the evolving needs of the industry and maintain our position at the forefront of facial recognition advancements.

For more information on our QAA product and its capabilities, please contact
News Articles