In recent years, face recognition technology has been widely used in the security field. With the continuous development of technology, it is getting closer and closer to our daily life, and mobile phones, shopping malls, parks, etc. can see it. According to relevant analysis reports, it is expected that the face recognition market will maintain a growth rate of around 20% this year. By 2022, the global face recognition market will reach $7.595 billion. It can be said that the next few years will be the key year for the mature and popular face recognition technology.
Face recognition, also called face recognition, is a biometric technology based on human facial feature information for identity recognition. It uses a camera or camera to capture images or video streams containing faces, and automatically detects and tracks faces in the images, and then performs a series of related techniques on the faces of the detected faces.
Domestic mainstream face recognition companies and characteristics
At present, there are many companies engaged in face recognition technology in the world. There are many well-known foreign companies such as google, facebook, stanford, carnegie, etc.; many domestic strengths have emerged in these years, such as Hikvision, Keda Xunfei, Shangtang Technology, cloud and technology.
These mainstream face recognition companies in China have their own technical solutions. For example, Hikvision's face recognition technology, its face recognition system is very good in recognition accuracy, front-end intelligence - squat, distributed intelligence - super brain NVR, central intelligence - Haikang face, all incorporate deep learning algorithm, Front and back depth intelligence provides powerful support for face big data applications.
Face recognition systems are widely used in safe cities, transportation hubs, smart healthcare, commercial chains, banks, schools, parks, border inspections and other industries.
The face recognition technology of Keda Xunfei has made a lot of breakthrough research on the accuracy rate. Its offline face recognition and video stream detection technology is also quite bright. Offline face recognition is further positioned in the detected face frame. The facial features and contour position of the face are currently positioned at the key point of the 21-point model. The average accuracy of the face key points is as high as 96%; the video stream face detection and key point detection, and the face detection in the static picture support offline use. , excellent performance in improving response time, this technology can be used in areas such as photo editing.
Its face recognition technology is widely used in many fields such as attendance systems, remote authentication, access control systems, and entertainment applications.
The face recognition technology of Shangtang Technology is very good in the calibration of video stream face, and the accuracy and delay are very good. Face recognition algorithm is also one of its advantages. According to the OFweek artificial intelligence network, Shangtang Technology once in the international authoritative face database LFW, its face recognition accuracy rate exceeded the human eye for the first time. For the security field, it launched the static face comparison system SenseTotem (totem), and dynamic face comparison - SenseFace. The totem is a set of map search system. It can capture the face snapshots in the surveillance video and compare the standard face photos in the search target library to help the police quickly confirm the identity of the suspects involved. SenseFace is used at the airport. Real-time large-capacity face recognition in large-scale video surveillance systems in public places such as train stations can provide real-time capture of face, attribute recognition, and key personnel track restoration in surveillance video.
In addition to its application in financial, access control, driverless, and other popular applications such as small coffee show, panda live, and other popular AR effects, face recognition technology comes from the technical support provided by Shang Tang.
Cloud from technology's face recognition technology is very distinctive, and its face recognition technology uses structured light-based anti-attack detection, which has a good effect in improving the accuracy. The 3D face recognition system of structured light technology has made a revolutionary breakthrough in accuracy, response speed and living body detection. 3D structured light face recognition technology can be widely applied in Internet of Things, mobile internet, banking, security, transportation, etc. Each scene that requires face recognition can better improve the attack prevention effect and comprehensively improve the accuracy of face recognition. Combining the latest algorithm from the cloud, it can achieve an accuracy of over 99% at a rate of one millionth. rate.
Its face recognition technology is widely used in banks, public security systems, shopping malls, etc., and establishes joint laboratories with the Ministry of Public Security, state-owned large commercial banks, Zhengtong shares, and the Civil Aviation Administration to promote the establishment of artificial intelligence product standards.
The current domestic face recognition technology faces the pain points in the market
Security, banking, and public security systems are the most widely used in areas where face recognition is currently the most popular. Future faces can shine in many areas, depending on technological breakthroughs and innovations. As far as the current status quo in China is concerned, face recognition faces two major pain points in the market.
First, the business model needs to be improved
The popularity of a new technology is related to many factors. As a company, cost control and profitability are its purpose. The face recognition technology has high threshold and high investment, and the application scenarios in the short term are very limited. Therefore, many enterprises are hesitant in the layout. They want to make face recognition technology enter the home like online shopping, a good business model. It is vital.
In the popularization of face recognition technology, many domestic companies have also made a lot of efforts. For example, Baidu has launched a product called Face Excellent, which is a tentative application of Baidu in artificial intelligence technology. Some people think that the launch of this application allows The mobile APP has broken through from the "model era" to the "surgical era." Because the application is connected not only to the user and the future business model, but also to connect Baidu's artificial intelligence and face recognition technologies, and has a high enough technical threshold to achieve difficulty. Internet giants such as Alibaba and Tencent have also tried commercial realization based on face recognition technology, but the results have not reached expectations.
Second, the current face recognition technology still has room for improvement
The current face recognition system mainly includes four components: face image acquisition and detection, face image preprocessing, face image feature extraction, and matching and recognition. As the data grows in the future, there will be even greater challenges in data matching.
Face recognition is very useful, but currently there are many deficiencies in face recognition technology, such as sensitivity to the surrounding light environment, which may affect the accuracy of recognition. In the face of dark, similar and other complex scenes, the recognition rate does not perform well.
In recent years, although domestic enterprises have achieved certain results in face recognition technology, there are still many aspects that need to be worked hard. For example, the recognition of faces, nighttime recognition, and infrared recognition through window glass is still waiting for enterprises to develop.
Where can the current face recognition technology be improved?
To make progress in face recognition technology, it is necessary to seek breakthroughs from several key technical points. Several key technologies for face recognition include:
Feature-based face detection technology—face detection by using color, outline, texture, structure, or histogram features.
Based on template matching face detection technology - extracting face templates from the database, and then adopting a certain template matching strategy, so that the captured face image matches the image extracted from the template library, and the correlation level and the matching template size Determine face size and location information.
Statistical face detection technology—by using a large number of face positive and negative sample libraries for “face†and “non-human face†images, statistical methods are used to train the system to achieve face and non-face The pattern of the face is detected and classified.
From a technical perspective, the approach to face recognition improvement can be derived from the above three aspects. Any breakthrough in detail technology can lead to improved recognition rates.
In addition to the above three technical aspects to improve face recognition, there are also many upstream companies doing research on the improvement of face recognition technology, such as improving the face recognition accuracy through improvements in components.
Infrared fill light sources for the user's face or eyes need to be bright enough and uniform, which is especially important for face recognition and eye tracking systems. Osram's new Synios P2720 uses infrared light with a wavelength of 940 nm to reduce red bursts. Previously infrared cameras had the best sensitivity to 850 nm infrared sources.
It is understood that this IRED is designed for 2D face recognition and is the latest addition to OSRAM Opto Semiconductors' existing line of biometric products.
The main research direction and development prospects of face recognition
With the advent of big data and sharing era, data security issues have become more and more important, and a new generation of technological revolution represented by face recognition has begun. These requirements for technology are getting higher and higher, requiring both data accuracy and data security. Face recognition is promising in this respect. As the main force of the industry, the company's technical strength and innovation ability determine the entire industry. The direction, any little technological innovation can bring about changes in the industry.
The main research direction of future face recognition will focus on some problems currently faced, such as the similarity of face and face structure, face posture, age change, lighting changes in complex environments, and occlusion of face ornaments.
The development of face recognition is not only limited by its own technology, but also closely related to the entire industry. In order to make artificial intelligence popular, many companies are actively deploying face recognition, including BAT Internet giants and companies that improve optics for face recognition. Ma Yun invested in face recognition technology company defiance technology and Shang Tang technology; Tencent established Utopia Lab, focusing on image processing, deep learning and other areas to carry out technology research and development; Baidu also set up a face recognition team to develop core technology. Recently, optical giant Osram acquired US company Vixar Inc. to strengthen 3D facial recognition skills and actively improve face recognition technology; by introducing Vixar technology in vertical cavity surface emitting laser (VCSEL), OSRAM will master more including ultra-precision 3D facials. Identifyed security technologies that can be used not only to unlock smartphones and consumer electronics, but also in industries that require higher security access control.
Relying on the rapid advancement of the Internet of Things and artificial intelligence, the face recognition application scene will become more and more extensive. With the R&D investment of national scientific research institutions, the research of technology on the enterprise, and the promotion of the market, these will be the signs of the bright future of face recognition. According to industry analysts, face recognition may become the mainstream of effective identification in the future. At that time, face recognition is not a new word.
Optical Rotary Sensor,Custom Encoder,Optical Encoder 6Mm Shaft,Handwheel Pulse Generator
Jilin Lander Intelligent Technology Co., Ltd , https://www.jllandertech.com