Ping An Technology developing AI face recognition technology with record results

March 15, 2017 - 

Ping An Technology Co. Ltd. is reporting key breakthroughs in the development of its artificial intelligence face recognition technology.

The company recently achieved a high score in the latest tests conducted by the University of Massachusetts’ facial recognition technology benchmark Labeled Faces in the Wild (LFW), scoring a face recognition ratio of 0.9960+/-0.0031.

The record results are slightly higher than that of other face recognition developers, which have published test results confirming an accuracy rate of approximately 99.5 percent, and significantly higher than human eye recognition’s 97.53 percent accuracy.

The technology was developed by researchers at Ping An’s Artificial Intelligence Laboratory, which was established in 2014 to study face recognition algorithms and the related core technology, and the company’s data platform development team.

The main focus of the Artificial Intelligence Laboratory is to increase its understanding of AI facial recognition technology’s potential.

The team is also responsible for the company’s business strategy, which includes the development of artificial vision, addressing issues related to natural language use in machines, and the research and development of artificial awareness towards unfamiliar objects and environments.

Within a year, the AI lab successfully worked through the stages of core technology research, product development and engineering, testing and implementation, and integration of the technology at a business level.

The technology has already been used in 108 different settings and scenarios, where it has facially recognized more than 103 million users.

Liu Fei, chief scientist of AI at Ping An Technology, said that the face does not contain stable biological characteristics, causing factors such as age and makeup to impact the accuracy and precision of face recognition.

Leveraging the data Ping An systems have extracted, the self-optimizing framework of the Ping An face recognition technology can now handle the effects of ageing.

Provided that the system has sufficient photographical evidence, even identical twins can be individually identified.

Due to the failure of industry regulations to keep up, the constraints of traditional norms and standards, as well as artificial intelligence itself being a relatively new concept, many of the existing technological advancements have still not been made available for release.

Scientists at Ping An AI Technology Laboratory are still working on decreasing the time required to complete the facial recognition process of human micro-expressions down to 30 milliseconds, while developing the emotional judgment and decision making ability of AI, Liu said.

“There is still a long way to go, but the technology has progressed significantly over the past few years. It is still being developed, but the development process is progressing quickly,” Liu said.

In addition to analyzing face recognition algorithms, Ping An Technology has also developed the capability to mine huge amounts of data.

Ping An Group has all the relevant financial industry licenses, and has developed several technological products.

Its AI technology has already been used in several areas of the financial industry including credit reporting, risk control, fraud identification, market investment analysis, and precision marketing.

Aside from the finance sector, AI technology can also be applied to help in providing accurate medical treatment, calculate personalized insurance premiums, and the field of graphics and imagery, among others.

Liu forecasts AI technology to have a significant impact across all sectors beginning with the financial industry.

“Artificial intelligence will create a phenomenon of change and development across all industries, but thus far we have only seen the tip of the iceberg,” Liu said. “Before the existence of artificial intelligence, we were unable to develop some aspects of the online financial industry. Now we possess this technology, we can develop these aspects in a safe, secure, and controlled environment. This transformation should help industries to not only exist, but thrive and flourish for many decades to come.”

Leave a Comment

comments

About Justin Lee

Justin Lee has been a contributor with Biometric Update since 2014. Previously, he was a staff writer for web hosting magazine and website, theWHIR. For more than a decade, Justin has written for various publications on issues relating to technology, arts and culture, and entertainment. Follow him on Twitter @BiometricJustin.