Новости компании

Closed-circuit television is already a familiar city life facility. Public squares, streets and lanes, and bus systems are all equipped with cameras in the name of fighting crime. The number of cameras in big cities has reached tens of thousands, generating massive amounts of video, most of which are daily clips, which are quiet and unattended.

The latest developments in the field of machine learning have spawned software that can recognize the unique features of a person's face through images or videos, with much higher accuracy than previous technologies. This also allows the camera to automatically track a large number of individuals in a crowd walking in public. If this task is done manually, the cost will be astronomical. The recognition software converts the face into a barcode. The camera becomes a scanner, automatically registering people who pass by the camera.

The police have not yet widely used the latest technology, and the accuracy of its Facial Recognition System is really flattering. The private group "Big Brother Observation" has investigated the use of facial recognition by South Wales police in public activities in the past two years. The agency found that its error rate was as high as 91%, which means that only 9 out of 100 people who are marked are actually the objects that the system hopes to mark.

This situation will end soon. Global closed-circuit television and police cameras can capture higher resolution images with simple upgrades. Such images help facial recognition algorithms work more efficiently. The camera system can be designed according to the needs of facial recognition, and use two or more cameras to shoot faces from different angles, so that individuals become easier to recognize. Silicon Valley uses powerful computers and large facial data sets to train high-precision software in the field of Face Recognition Device. It has just begun to set foot in the security market, but it will usher in accelerated development in 2019.

Face Recognition Device

The impact is really worrying. As countries intend to control the infrastructure for recording facial images in public, accurate facial recognition is likely to be improperly used. For example, government authorities may capture the "face prints" of a group of protesters, then arrest any protester who may have a criminal record, and then directly include the rest on the list of suspect faces. Private companies may also abuse facial recognition technology. Shopping malls can use this to create a shared customer database and a database of products viewed by customers, similar to today's online tracking, which is not conducive to customer privacy protection.

Facial recognition may also link the digital world with traces in real life. It converts facial features into an address, and associates real-world behavior with online data, and vice versa. The combination of physical indicators such as web tracking and facial recognition will mean that the space where humans are not tracked will shrink in 2019.

It is difficult to predict the specific negative effects that new technologies will bring. In the early days of online advertising, few people would have thought that it gave birth to a network dedicated to eyeballs, and in this network, those insignificant anecdotes spread farther and more influential than informative and accurate events. . We have reason to be concerned about the principles that facial recognition may break and its potential destructiveness.

Humans need a space where they are not tracked, where they can freely meet people they want to see. When the world is full of ubiquitous Facial Recognition, whether it is a coffee party, a midnight walk, or shopping in a mall, everyone will be assigned their own facial prints and identities. Strong laws protecting individual rights will be the best hope for restricting such tracking. Although we have opened the Pandora's box of facial recognition, there are still opportunities to limit its negative effects.