The powerful symbiosis between biometric and video analytics technologies has become an increasingly crucial factor in law enforcement in recent years, particularly as as CCTV and smartphone culture permeates in our "always on" culture.The potential for wider applications in security has expanded greatly since such tools first came to wider public attention following the Boston Marathon bombing in 2013 – but users also expect seamless real-time face recognition for fun or time-saving tasks related to retail and travel. However, while the potential is indoubtably there to save hundreds of thousands of man hours, to track people globally and to save lives and prevent crimes, there are also a whole host of challenges raised by video analytics, such as how to identify faces from innumerate different positions, and in tricky lighting and speed scenarios.Planet Biometrics talked to Elke Oberg, Marketing Manager, about how that company is addressing these challenges with innovative new solutions.What led Cognitec to focus on video inspection tools?Cognitec has been offering pioneering technology for real-time face recognition in video streams for almost a decade. The path of continued innovation naturally led to extending the same capabilities to the investigation of recorded video material. With widespread ownership of mobile devices, and higher quality of CCTV footage, law enforcement agencies are collecting more video material than ever to find persons connected to crimes. It is extremely time- and staff-consuming to watch hours of video material, hoping to spot suspects, and evaluate their likeness. Using face recognition technology brings a new level of search automation and efficiency to the investigative workflow.Can you give us an overview of the new features you have been working on?Our latest product release, FaceVACS-VideoScan LE, enables fast import of video footage and detailed examination of the extracted facial images. Agents can find known or unknown persons in multiple videos to quickly narrow down suspect investigations. The search could reveal, for example, if a suspect was seen in various locations within a set time window, always with the same group of persons, or in one location too many times during the day.Users can view each appearance of a person in form of a video snippet. From the video snippet, they can instantly switch to the original video material and view the corresponding scene. A click on another face within the video frame shows all appearances of this person, who might in turn be connected to a suspect or crime scene.How important is face recognition for the future of search automation?Image and video material from crime or terrorist scenes has become a primary source of evidence along the investigation trail. Security staff and law enforcement agencies will need reliable tools to analyze the material and quickly utilize the search results. Immigration and border control procedures will need fast automated searches to check vast databases of images to either find known suspects, or to check a person's visa status. In refugee situations, where people often enter the country without valid ID documents, governments are now relying on face recognition to search databases for previous records and detect fraudulent activities.How do you think the biometric exit situation will play out in the US?At the moment, many stakeholders are evaluating the technical feasibility and costs of such a large-scale system. Face recognition seems to win the race for biometric of choice, with digital ID photos already stored in most passports, and the possibility of touchless interaction with a camera. Systems of such magnitude will deliver previously unknown challenges, with data security the most urgent to address. On the more practical end, it seems still quite unclear who will be responsible for the inspection at which point of the airport journey, and what will happen with visa offenders. Self-service terminals would be the preferred mechanism for the travelers, but the complex, costly tasks are connected to their implementation and supervision.What important trends in face recognition are impacting on the industry?I see three important trends that are currently shaping our industry. First, commercial giants, with gigantic facial image databases, like Apple, Google and Facebook, are driving research, development and public acceptance. And acquisitions continually re-arrange the leaderboard of the biometrics industry and create extremely resourceful pools of innovation. Also, certain governments are providing face recognition startups with limitless funding to develop powerful technologies for widespread surveillance. These companies are now winning vendor tests and offering their products to the international market.Second, the debate around privacy and data protection continuously heats up most around face recognition use cases. Where does face recognition make sense, where does it truly enrich or simplify our life? Can we protect and contain such sensitive data? Do we really need digital signs that show targeted advertising based on gender and age estimation? Does face recognition in a train station rob us of the freedom to move about anonymously in a public space, while possibly providing zero crime-fighting benefit? In democratic countries, the public, the consumers, the consensus shall decide. And the industry will listen and adapt.However, and this is the third and ongoing trend, customer demand for reliable face recognition products continues to increase. The technology has conquered the consumer space, border control agencies are using face recognition as the primary biometric technology, and casino operators rely on it more and more to protect their premises and clients. Promising times for our industry!
Interview: Cognitec on the importance of analytics
