Published on February 6th, 2020 | by Emergent Enterprise0
AI Facial Recognition and IP Surveillance for Smart Retail, Banking, and the Enterprise
We already know that we being captured on video almost everywhere we go. But, we are not only being archived, there is a good chance we are being recognized as Susan Fourtané reports at Interesting Engineering. Through a combination of facial recognition technology and artificial intelligence cameras can spot a face and match them to faces in a large database. This technology is ideal for security reasons at large events and banks and other places where criminals might be spotted. But, it also raises many questions about personal freedoms and ethics in using the technology.
With the Facial Recognition market worth $7.0 billion by 2024, retail, banking, and the enterprise are rapidly adopting the technology coupled now with IP surveillance.
Facial Recognition technology detects faces in the camera’s field of view and matches them against faces previously stored in a database. Anti-spoofing is provided through liveness testing without the need for a stereo or a 3D camera. Face Recognition technology is now taking a further step as it is being combined with IP surveillance.
Gemalto, a part of the Thales Group and a company that focuses on Digital Identification and Data Protection in order to counter the two root causes of cyberattacks, identity theft, and unencrypted data, defines Facial Recognition as the process of identifying or verifying the identity of a person using their face. It is a technology that captures, analyzes, and compares patterns based on the person’s facial details. The face detection process is a basic and essential step allowing the systems to detect and locate human faces in a set of images and videos.
The face capture process transforms the analog information contained by a face into a set of digital information based on the person’s unique facial features. By using this data, the face match process verifies if two faces belong to the same person. Face Recognition is considered to be the most natural of all biometric measurements currently in use today.
Facial Recognition is set to be a major topic for the 2020 Summer Olympic and Paralympic Games in Tokyo, Japan, which are taking place from July 22 to August 9 and from August 25 to September 6 respectively, as the first time the technology is being used in the Olympics. Facial Recognition technology will be used in Tokyo 2020 to identify authorized persons and grant them access automatically, enhancing their experience and safety.
The technology is clearly having a momentum. According to BBC News, U.S. supermarket giant Walmart has confirmed it uses image recognition cameras at check-outs to detect theft. Having used the technology already in more than 1,000 stores, the company has said that it had made “an investment to ensure the safety of our customers and associates.”
Although it could be argued that an ideal environment, such as an airport check-in, where the face is straight on and well lit, and the camera is high-quality, AI-powered face recognition is said to have become now better than human; and this has been this way since at least 2014.
Rise of Facial Recognition for IP surveillance
As advances and innovation around Facial Recognition technology continues to evolve even more, one of the latest trends come from CyberLink’s FaceMe® AI Facial Recognition engine integrated into Vivotek‘s IP surveillance solutions of network cameras and back-end video management software. This integration enable security operators to receive accurate Facial Recognition alerts based on both blacklists and whitelists.
According to Dr. Jau Huang, CyberLink’s Founder and CEO, “the demand for Facial Recognition is booming, driven by the latest IoT and AIoT innovations, and are enabling a wide array of scenarios across industries such as security, home, public safety, retail, banking, and more.” He says that each application is dependent on the performance of the cameras used to capture faces and by integrating FaceMe into Vivotek’s surveillance devices it is possible to bring accurate and reliable new solutions into the market.
Powered by Deep Learning and Neural Network algorithms, Cyberlink FaceMe is one of the most accurate AI Facial Recognition engines, according to the company. However, this statement is totally backed up by a recent Face Recognition Vendor Test (FRVT) conducted by the U.S. National Institite of Standards and Technology NIST, CyberLink ranked 12th among all participants in FRVT 1:1 (WILD 1E-4), confirming that FaceMe is a world-leading facial recognition engine. The NIST report (PDF) details recognition accuracy for 127 algorithms and associates performance with participant names.
The NIST FRVT WILD 1E-4 dataset consists of faces extracted from surveillance camera footage or photos, encompassing a wide array of real-world situations including a range of capture angles, poor lighting, or partially covered faces. The image variability simulates real-world use cases where systems would be required to accurately identify individuals in multiple different settings.
Some applications for this technology include the retail industry, banking, organizations transitioning into digital transformation, and those companies that want to make their offices smarter.