Face Recognition

FACE RECOGNITION

Augmenting

Intelligence

NtechLab research team is a group of experts in the field of deep learning and artificial intelligence. We like to invent algorithms which work in unconstrained real-life scenarios

Download PDF

Our algorithms

Face recognition for verification

The best verification algorithm in the world. Proven by the latest independent testing of face recognition engines.

Winner of the Intelligence Advanced Research Projects Activity (IARPA) Face Recognition Prize Challenge (FRPC). See the report below and a detailed report on nist.gov.
At the recent Face Recognition Vendor Test, conducted by the National Institute of Standards and Technology (NIST), NtechLab’s face recognition technology showed the best performance on sophisticated datasets – “wild” and “child” – among all participants for large-scale verification.

Face recognition for identification

The IARPA Face Recognition Prize Challenge found NtechLab to have the world’s fastest algorithm. It is currently the world’s only algorithm with sub-linear search time.

NtechLab’s algorithm works on planetary scale datasets with billions of faces, showing the best accuracy and speed performance:

Search time on huge datasets:

250 million faces in dataset – search time is less than 0.2 sec.
500 million faces – less than 0.3 sec.
1 billion faces – less than 0.5 sec.

Face detection algoritms

Our face detection algorithm is renowned for revolutionary speed and accuracy. In a single shot our algorithm processes an unlimited number of faces, perfectly tailored for mass events. It is resistant to lighting conditions, posture, head poses and tilts.

Age and gender recognition algorithm

Detects a person’s age with 95% accuracy in 5 years’ intervals.

Detects a person’s gender with 99% accuracy.

Age and gender detection can be adopted across a wide range of use cases and markets including targeted offline advertisement, access control and data enrichment.

Emotion recognition algorithm

Detects 7 primary and 50 compound emotions of a person. Captures and interprets emotional data to deliver actionable insights.

Emotion detection can be adopted across a wide range of use cases and markets including entertainment and media, client satisfaction measurement.

Algorithm’s quality is proven by 1st place at EmotioNet Challenge 2017.

Our products

FindFace Security

FindFace Security provides intelligent video analytics employing facial recognition technology. The product detects faces in live video streams, checks these against monitoring databases then notifies the relevant bodies upon there being a positive match.
An exact facial recognition algorithm and fast implementation ensure this is the ideal solution for corporate and governmental security needs. One distinctive feature of FindFace Security is its ability to support an infinite number of video streams and facial database entries, ensuring the system offers unlimited scalability. Using the intelligent criteria-based facial search, a suspect can be filmed, detected and tracked in a split second. A user-friendly GUI ensures parameters can be swiftly set up in just a few click.
Fine-tune notifications to receive only those needed due to the capability to manage user roles. FindFace Security easily integrates with a range of external systems and services, including access control, POS, video monitoring and many others.

The product can be applied to a variety of fields ranging from metropolitan public security through to counting the number of visitors to retail stores.

FindFace SDK

The FindFace SDK is a dedicated C Library that leverages advanced facial recognition on neural networks.
The FindFace SDK offers fast and accurate results regarding three basic facial recognition tasks: detection, extraction of biometric template, and verification. A great benefit of the FindFace SDK is its capability to process not only video feeds but also still frames that render facial recognition a viable option for environments in which no cameras and/or video processing hardware can be installed.
Written in C/C++, the Library supports multithreading and easily integrates into other products and services.

The FindFace SDK successfully can be embedded into e-gates, self-service kiosks and wearable devices. The product application ranges from airport check-ins to corporate access control.