Facial recognition technology or usher in the "second spring"

The rapid popularization of smart mobile devices and the continuous improvement of shooting functions have created numerous new application scenarios for facial recognition technology and are giving it new life.

Strictly speaking, facial recognition technology is not a very novel technology. Its history can be traced back to the 1960s. In the field of security, facial recognition technology has been used for many decades. However, this technology has not been able to flourish in the past half century. Ordinary people know more about it from Hollywood movies. FBI and CIA use advanced face recognition through cameras in all corners of the globe. The system can lock the suspect's position within a short period of time. Of course, these are more of an exaggeration of art. In the real world, facial recognition technology is far from as cool as it is on the big screen.

Over the years, the application of facial recognition technology has been “trapped” in the security field, making it difficult to break through. However, with the intervention of Internet and IT companies in recent years, this technology is ushering in its "second spring".

At the beginning of June, Sony released the facial recognition software SOEmote at the E3 game conference. It can capture the player's facial expressions in real time through the webcam and feed back to the player's face at the same time; on June 18, Facebook was completed at a price of about 60 million US dollars. The face recognition company of Israel Face. The acquisition of com. Prior to this, Apple and Google acquired facial recognition technology companies Polar Rose and Pitt Patt in 2010 and 2011, respectively.

In China, on June 4th, NetEase began to test the face recognition system of NetEase's mailbox. During the same period, Shanda also released App “Smart Album” based on face recognition technology.

In just six months, a number of IT internet companies have launched products and services related to facial recognition technology at home and abroad.

Untimely

Face recognition technology is a kind of biometric recognition technology based on human facial feature information. It first uses the camera to capture images or videos containing faces, and then uses computer image processing techniques to extract portrait feature points from the screen and uses biometrics principles to analyze the modeling, that is, face feature templates. At the end, feature analysis was performed on the facial image of the subject to identify the identity.

Research on facial recognition began in the 1960s, but in the first 30 years of its existence, this technology has not produced significant results except for some basic research. After the 1990s, with the rapid development of computer technology, facial recognition technology finally ushered in the spring. Throughout the 1990s, there were many theoretical breakthroughs in facial recognition technology, and the US military began to fund related research for the needs of anti-terrorism, which also enabled the rapid transfer of face recognition technology to practical applications. The FEERT project funded by the US Department of Defense organized three facial recognition evaluations in 1994, 1995, and 1996. Several well-known facial recognition algorithms all participated in the test. The three evaluations directly promoted the improvement of facial recognition algorithms.

After the "September 11" incident, in order to curb terrorist attacks, the United States paid more attention to facial recognition technology and was more widely used in the field of security.

After experiencing a major outbreak in the 1990s, facial recognition technology has encountered bottlenecks in the last decade or more: not only has the technology itself not been a substantial breakthrough, but more importantly, it has been limited to security. There has been no revolutionary product. Although the use of facial recognition technology products has begun to appear in the civilian market in recent years, it is still concentrated on access control, attendance, and monitoring systems, and it is still unable to escape the shackles of security. The only innovation comes from the digital camera industry. In 2006, Nikon took the lead in applying facial recognition technology to its products. Cameras that use facial recognition technology can automatically search for human faces and prioritize focus when shooting. Under the leadership of Nikon, facial recognition technology has also been widely adopted by other manufacturers, becoming the standard for digital cameras.

Although the camera industry introduced face recognition technology, it did not bring any fundamental change to the technology and camera industry. It only improved the user experience to a certain extent. For the average consumer, the built-in camera face The recognition function just makes the photos of the portraits clearer and nothing more.

However, the involvement of the Internet and IT companies has changed the fate of facial recognition technology.

New Life in the Mobile Internet Era

The Internet giant’s acquisition of facial recognition technology companies in the last two years is reviving facial recognition technology. The rapid spread of smart mobile devices in the past few years has brought new opportunities for facial recognition technology. The increasing number of pixel cameras on mobile phones and computers has helped facial recognition technology to open up new opportunities in addition to traditional security services. Application scenario.

At the end of 2010, Facebook was the first in the United States to introduce “Tag Suggestions” using face recognition technology. In June 2011, Facebook expanded this feature to most countries outside the United States. In response, Google also launched a similar feature "Find My Face" on Google+ in December 2011, allowing users to automatically find their own images in their friends' albums through facial recognition.

Facebook just acquired by Facebook. Com also recently launched KLIK, a mobile camera application that uses face recognition technology in the App Store. The main application of this application is not the rich filter function, but the social sharing function. It can automatically identify the subject's identity from the screen during the photographing process, and quickly share the user's social networking site accounts such as Facebook. It is also the improvement of camera functionality that Internet companies and camera manufacturers present completely different ideas.

Three months ago, Facebook’s acquisition of Instagram was a sensation in the industry, from Instagram to Face. Com, Facebook's ambitions in photo socialism are evident. The launch of KLIK should only be a test of water. It can be predicted that Facebook will definitely integrate KLIK and Instagram in the future.

Face. Com is also exploring how to build a viable business model based on face recognition. In March of this year, Face. Com announced the addition of a new age recognition feature to the API, which can estimate the approximate age of people in the photos, which will help improve the accuracy of advertising. IT giant Intel also has a similar plan, according to Reuters news, Intel is integrating facial recognition technology into TV set-top boxes, enabling set-top boxes to recognize the gender and age of the audience before the TV, thus providing an important reference for advertising. According to reports, Intel's set-top box products are expected to be available at the end of this year.

The popularity of facial recognition technology also involves an old-fashioned but unavoidable issue - the protection of user privacy. Radical Facebook has already had a lawsuit in Germany. In August last year, the Hamburg Data Protection and Information Security Agency ruled that Facebook’s facial recognition technology violated European and German data protection laws and ordered Facebook to delete relevant data; while cautious Google has already Being careful everywhere, he still got into trouble in France. In March this year, the French National Freedom of Information Committee challenged a series of Google services including facial recognition.

In addition to the ethical controversy, facial recognition technology is still unsatisfactory at the technical level, especially on non-professional equipment. Last year Google released the Android 4.0 system with face recognition technology, but soon after the release, users have taken photos with the system. Deng Yafeng introduced that although many Internet companies are launching security services based on face recognition, the recognition rate is still not high enough from the current level of technology. It can only be used as an auxiliary measure and cannot be a single factor.

In today's era of big data, facial recognition technology wants to go further, so there are many difficulties that need to be overcome, both technically and ethically.