As one of the most representative technologies in artificial intelligence, face recognition is playing an increasingly important role. In China, “facial payment” was even selected by MIT Technology Review as one of the “Top 10 Breakthrough Technologies of 2017,” making it the first Chinese technological breakthrough to enter the list since its inception 16 years ago.
However, as face recognition technology becomes more deeply embedded in real-world applications, it faces growing challenges—foremost among them is anti-spoofing. For example, during this year’s CCTV 315 investigative program, a face-swapping app easily unlocked facial recognition applications. Recently, a flagship smartphone from an international tech giant had its much-touted facial recognition feature easily bypassed by a static photo. These incidents highlight the fact that, beyond high accuracy, security concerns are drawing increasing attention. To ensure the security of face recognition, liveness detection has become crucial.
What is face liveness detection? Simply put, it’s the process of confirming during face recognition that the subject is a real, live person—not a photo, video, or other spoofing medium.
Looking at most face recognition systems in the market today, those that rely on 2D facial recognition are vulnerable to spoofing via photographs or video clips. Vizum Intelligent applies its independently developed stereo intelligent vision technology to 3D face liveness detection, upgrading facial recognition from 2D to 3D and effectively addressing the security loophole of photo-based attacks.
As one of the core technological applications of Beijing Vizum Intelligent, the newly released face liveness detection SDK uses an innovative binocular algorithm to calculate high-precision facial data. It precisely identifies and locates key facial features such as eyes, ears, and nose, and computes various spatial metrics such as inter-pupillary distance, nose height, and distances from eyes to mouth to ears—with a margin of error under 1mm. This enables accurate detection of real, live faces and effectively prevents spoofing via flat photos, bent or manipulated images, Photoshop edits, and video replays.
Compared to most monocular 2D visual technologies on the market, Vizum’s solution offers higher detection accuracy, faster processing speed, and less need for database maintenance and updates. It also features adaptive learning capabilities that help overcome challenges posed by changing lighting and complex background environments.
At present, Vizum has made its stereo liveness detection technology available for integration in various sectors such as finance, security, education, and social services. The company provides an SDK that can be quickly integrated into customized devices running on Windows 7, Windows 10, Linux, and other platforms.
As an innovator in stereo intelligent vision, Vizum is committed to equipping machines with human-like “intelligent eyes” through this technology, ultimately realizing true artificial intelligence. The success of our 3D face liveness detection technology once again affirms the validity of our development philosophy of intelligent innovation. For specific applications, we are also capable of rapid development and deployment, speeding up product implementation.
Looking ahead, we will continue to utilize high-precision facial feature data to advance stereo face recognition technology and accelerate its commercial application. If you are interested in our products, please contact us right away (Mr. Zhang: 010-82090180). Vizum looks forward to joining hands with you to build the strongest defense in face recognition and safeguard the security of intelligent living.