Prof. Ma Lizhuang’s Research Team Broke 18 World Records

International Affairs Division 2021-06-09 1024

As the most populous country with a high mobility amid the rampant pandemic, China has witnessed the burgeoning of various “health codes” and almost every Chinese has used them in one situation or another.

The research team led by Prof. Ma Lizhuang from SJTU School of Electronic Information and Electrical Engineering is the facilitator of this phenomenal technology. The project they carried out, titled “Technology and Application of Character Visual Recognition for Complex Scenes”, has broken 18 world records and was awarded the Special Prize of Shanghai Science and Technology Advancement Awards conferred on May 25th.

 

In the wake of the epidemic, Prof. Ma’s team quickly responded to the need of epidemic prevention and control and created a 3D multi-modal facial verification solution based on leading face recognition and security technology in the world. The core of health codes” is remote identity verification based on light reflection, which helps to ensure the identity authenticity. When a user logs in, the light reflection examination will be triggered to collect real-time data and compare it with the face recognition result. Only by passing this test will the user be given access to the instant “health code”.

When the epidemic was brought under control, the character visual recognition technology also contributed to the resumption of work and production through integrated use in various online and offline scenarios. Online remote identity verification has been applied in more than 140 industries to facilitate the transportation safety of billions of people and to build the new normalcy in epidemic prevention and control.

From intelligent finance to AI system for seeking lost relatives, character visual recognition technology has proved its applicability in a series of complex scenarios. In the year 2015, Ma’s team built the world’s first remote face-identity verification system for commercial banking system, which requires an error margin lower than one trillionth.

So far, this project has been granted,passed 1 IEEE international standard and 8 national standards and produced 290 high-quality research articles. Behind the spectacular achievement was 10-year’s commitment and devotion. A decade ago, when cloud computing and big data were the focus of the academia, Prof. Ma Lizhuang had been leading his team in improving the accuracy of image and geometric recognition modeling, a less explored filed.

 

 

Over the past ten years, the project has generated over 10 billion economic benefits and constantly advanced the progress of artificial intelligence related industries.

 

Baidu
map