According to the US The Intercept website, Morpho is working with IntelCenter to develop an Internet face recognition technology similar to "Google" to help law enforcement, intelligence and military agencies search for and identify terrorists.
Morpho is an internationally renowned provider of biometrics technology and has previously worked with the US Federal Bureau of Investigation to develop face recognition technology. The new technology developed in this time combines not only advanced face recognition technology, but also a face database containing terrorist data. IntelCenter is a well-known private intelligence company that provides services to government clients, including tracking information, videos, social media, hostage crisis events on the Internet, and information disseminated by terrorist organizations and their supporters
Some people questioned the accuracy of the technology, saying that Morpho's face recognition technology used by the US Federal Bureau of Investigation is only 86% accurate, which means that 14 faces are recognized for every 100 faces. Is the result of misidentification. In fact, the performance of face recognition technology is affected by many factors, such as the size of the face database and the quality of the face image. Whether the face is extracted from a poorly lit video or extracted from a clear driver's license photo, it is difficult to correctly identify the length of the nose or the distance between the eyes.
It is reported that the US military has a long history of using biometrics (such as face recognition and fingerprint recognition) technology to identify terrorists. US soldiers in Afghanistan have developed handheld biometric devices that match targets based on features such as iris, fingerprints and faces. There is also an agency within the US military called the Defense Forensics and Biometrics Agency that specializes in this area. With the continuous development of counter-terrorism actions, biometrics technology has gradually been used in law enforcement operations in various states and localities in the United States. Morpho's cooperation with IntelCenter is just one example.
With the rapid deployment and application of video surveillance cameras, millions of face and video images are produced every day in the world. How to quickly and accurately search these images to identify and identify suspects or terrorists is a difficult and complex task. Michigan State University developed a large-scale face recognition technology at the end of 2015. The technology uses video surveillance images or crime scene images to quickly search large-scale face databases and find suspects by getting the best matching search results list. This technology can process and search for unconstrained images of changes in poses, lighting, and background, and if it is integrated with NEC's Neoface algorithm, the correct recognition rate will be higher. The researchers used the 5 million real images of the Boston Marathon bombing to experiment and pinpoint the suspect.
Leading company in video surveillance technology - Hitachi, Japan launched the latest video analysis system called "Hitachi Video Analysis (HVA)". The system combines computer vision technology with advanced analytics to provide users with a variety of services.
Operational and business intelligence: Provides reference for optimal management of city managers and airport managers through demographics, team testing, behavior analysis, flow direction statistics, and image enhancement.
Intelligent transportation, traffic and parking management: Provides vehicle counting, traffic analysis, parking space detection and license plate recognition.
Public safety: Through intrusion detection, remnant detection, facial recognition and other technologies to help the public safety department to protect the safety of citizens and property.
The Hitachi Video Analysis system consists of a target detector, a scene analyzer, a video enhancer, and a camera status monitor. The target detector can perform 4D reconstruction on 2D images to improve recognition accuracy; the scene analyzer ensures that the analysis engine under each analysis module uses a custom algorithm to improve analysis performance and accuracy; the video enhancer can filter those possibilities Interference factors that affect resolution, such as rain and snow; the camera status monitor can automatically detect tampering or damage to the camera and alert you when a problem occurs.
Article taken from: China Security Journal
Author: Yu Shi
English 
简体中文