Problems that cannot be solved by face recognition can be solved by "pedestrian feature recognition"

　　"Artificial intelligence is not mysterious. What matters most is how to use it to solve practical problems." Yan Zirong, CEO of Ansoft Technology, repeatedly raised this point of view in an interview with Yiou. Let's learn more about it with the editor of Mobile Phone Portable.

　　Over the past decade, with the construction of "safe cities" across the country, the security industry chain based on video surveillance has also developed rapidly. Shenzhen Anruan Technology, which officially landed on the New Third Board in 2016, has been deeply engaged in the field of mobile policing for a long time. With the accumulation of experience and technological breakthroughs, Anruan Technology is also using technological innovation to solve the pain points of police security.

　　Traditional video surveillance systems have many pain points

　　The “Safe City” movement began to take off between 2004 and 2006, and subsequently, a large number of video surveillance projects were built across the country. Tens of thousands of cameras were installed throughout the city, providing powerful support for public security prevention and control, criminal case detection, and other work.

　　However, with the increase in the number of cameras and the resulting huge amount of information, it has become a burden in actual operations. For example, reviewing videos is particularly time-consuming and laborious. Many cities have installed a large number of cameras, but they are usually difficult to use and are only used when they are needed to solve a case. "It's like they grow a lot of crops, and what we do is process them into different foods (products) and then apply them to different scenarios, such as public security prevention and control, criminal investigation, and traffic management," said Yan Zirong.

　　Since its establishment in 2000, Ansoft Technology has been engaged in the research and development of products in the field of police security, such as mobile police equipment and systems, and has a deep understanding of police security work. In their view, traditional video surveillance systems have the following pain points:

　　1. Massive video information is rough and inaccurate, and it is urgent to refine "ore" into "gold";

　　2. Due to the large amount of video, reviewing is time-consuming and laborious, and manual methods are ineffective;

　　3. Massive stored videos cannot be called up at any time, which is not flexible enough;

　　4. The efficiency of the original system was not fully utilized, and 99% of the information was in a "dormant" state for a long time.

　　Yan Zirong, CEO of Anruan Technology, told Yiou that these pain points have existed for a long time. In the past, the technology was not mature enough to solve them. However, artificial intelligence has made rapid progress in the past two years and can be completely solved with artificial intelligence.

　　Using artificial intelligence to "activate" massive video information

　　The Security Industry Research and Analysis 2017 Report released by iResearch pointed out that AI empowers security by using computer vision recognition technology to solve the problem that massive video surveillance data cannot be well utilized. From massive data to information to intelligence and then to insight, it is necessary to extract the value of data layer by layer, and computer vision recognition technology can well structure what is "seen" into attribute information with "people", "cars" and "objects" as the main body.

　　Yan Zirong believes that many companies are now working on facial recognition , but in fact, the application scenarios are relatively limited. Take the video surveillance system of the safe city as an example. The images captured by ordinary cameras can hardly see the faces of people. Therefore, facial recognition plays a limited role in traffic management, public security prevention and control, and criminal investigation. He further mentioned that facial recognition is more used in surveillance. That is, if someone has been confirmed to have committed a crime, this person can be searched out through facial recognition, but it does not solve the problem of how to detect suspicious persons in advance.

　　The VPAIS system developed by Anruan focuses on proactively attacking in advance, detecting suspicious persons, and promptly handling possible dangers to "prevent problems before they happen." The VPAIS system uses artificial intelligence deep learning technology and big data technology, and uses edge computing, image structuring, image big data and other means to activate "sleeping" monitoring data, so that monitoring videos can be called, retrieved, analyzed, and applied at any time. In summary, it is to use artificial intelligence to upgrade the traditional monitoring system, and equip the "safe city" system with "optical nerves" and "super brains" to make it more intelligent.

　　Technologies used in the VPAIS system

　　Unlike face recognition, the VPAIS system uses self-developed "pedestrian feature recognition" technology, which includes pedestrians and vehicles. After classifying people, cars, electric vehicles, and bicycles, it specifically analyzes the unique attributes and behaviors of each individual. There are some special cases, such as asking the system to identify "people with knives". Under normal circumstances, surveillance videos rarely capture this situation, but such images can be actively input to allow the machine to learn what characteristics "people with knives" have and gain recognition capabilities.

　　By providing the VPAIS system to public security organs, it can help them improve the efficiency of public security prevention and control warning, improve the efficiency of criminal case detection, help manage traffic road conditions and ensure smooth roads.

　　Yan Zirong told Yiou that the VPAIS system has the following values: First, it is low cost. The usual practice is to replace the camera and match it with the supporting system, but the cost is high. The VPAIS system can be connected with the existing monitoring system of the public security organs, and there is no need to replace the smart camera, which is easy to transform; second, it is efficient. In the past, it took several days for several people to review a case, but now it can be solved in seconds; third, it reduces the work pressure of the police. Watching surveillance videos all day is a very painful experience. This greatly reduces the work intensity and saves unnecessary labor.

　　The technology itself is not that great, the key is where it is used

　　In terms of technology, the VPAIS system has already achieved a search of 1 billion images in seconds. Yan Zirong said, "Breaking through technology is important, but solving practical problems is more important. Artificial intelligence technology is not that high-end. The key is where to use the technology and what problems to solve. We did not build a new business system, but used artificial intelligence to solve existing problems that have not been solved."

　　Talking about the city's public security, Yan Zirong believes that among all the public security and criminal cases, major cases are a minority after all, while ordinary cases such as fighting, theft, robbery, and destruction of public property account for the majority of a city's public security. However, the number of these cases is relatively large, and they are precisely what makes citizens most dissatisfied and have the greatest impact on the overall public security of the city; the public security department should pay more attention to such cases in order to truly improve the level of public security management, improve the sense of security and happiness of residents, and at the same time prevent minor cases from happening.

　　For the public security organs, it used to take up a lot of resources to solve minor cases. The use of the VPAIS system can save police resources and make minor cases easier to solve. Ansoft Technology said that after they implemented the electric vehicle management system in Jingdezhen, Jiangxi Province using artificial intelligence + Internet of Things technology, the crime rate dropped by 60%. From the perspective of criminals, the "cost" of committing crimes has increased, and as long as they commit crimes, they will be caught immediately, so they have to restrain themselves.

　　The idea of ​​"finding problems - solving problems"

　　The realization of the value of artificial intelligence is inseparable from specific application scenarios and the resolution of specific problems. In terms of security innovation, it is more important to study and discover the pain points of the industry itself, and then find solutions to the pain points from artificial intelligence technology.

　　Ansoft Technology is based on mobile police systems. The reason why it has business innovation in the direction of artificial intelligence is mainly due to its understanding of the police security industry itself, identifying the most significant pain points in this industry, and finding solutions to the problems from technology. Its fundamental idea is "finding problems - solving problems", rather than imposing artificial intelligence on a certain scenario and then looking for needs. This idea is also in line with other security fields, and even other industries' use of artificial intelligence.

    The above is an introduction to the problem that cannot be solved by face recognition in portable mobile phones, and how it can be solved by "pedestrian feature recognition". If you want to know more relevant information, please pay more attention to eeworld. eeworld Electronic Engineering will provide you with more complete, detailed and updated information.