by Han-Ru Xie
With the development of AI technology, the accuracy of facial recognition systems has greatly increased. With new technology comes great convenience, but, inevitably, there will be problems. The following are a few advantages and hidden concerns in facial recognition technology.
Facial recognition that is efficient and widely used
Facial recognition is a type of biometric technology. Compared with fingerprint and retina scanning that have been widely used yet requires direct contact or close proximity for successful scan, the non-contact feature of facial recognition can bring higher efficiency for authentication.
Low cost is another advantage, as facial recognition works mainly on the images collected. Today, any smartphones can take pictures of high pixels comparable to those from monocular cameras, so it is not difficult to obtain high-quality images, which greatly reduces the cost of equipment. This technology features its timeliness and low cost, which is user-friendly, so it is suitable for daily use.
Facial recognition can be applied in vast areas, including access control, surveillance, finance, logistics, and even government law enforcement agencies. Even today, as the epidemic is spreading worldwide, this technology can be integrated with a body temperature measurement system for disease prevention.
Hidden concerns under the conveniences
As much as this technology makes our life more convenient, there are still concerns to be addressed. The biggest problem lies in the uncertainty of processing information and the invasion of privacy.
Unlike fingerprints or retina, human faces change with age or after plastic surgery, and, sometime, different lighting, angles, and expressions can affect the results of recognition.
There are examples where the Detroit Police Department caught the wrong person due to system misjudgment last year. According to The New York Times, the database of the facial recognition system in the U.S. collects data mostly from Caucasian male and white; therefore, there is a greater chance of incorrect matches when it comes to people of color.
In addition, using simply one photo, the system can collect information including a person’s identity, name, age, background, and phone number. Some information is even collected without anyone knowing it, such as where people have been to and what products they have purchased. Because of this, there are concerns regarding the invasion of privacy.
Apart from the above two issues, telling the difference between real human faces and those from photos as well as improvement of recognition efficiency are also challenges that need to be overcome.
Facial recognition brings a more convenient life, and it is also the general trend for future development of authentication systems. However, there are still room for improvement and technical as well as regulatory issues that need to be resolved. It seems that there is still a long way ahead before it is entirely widespread.