Language: Chinese line English

Access control news

Face Recognition access controller Industry Research Report

A few days ago, the "2020 Face Recognition Industry Research Report" initiated by the National Information Technology Standardization Technical Committee Biometric Recognition Sub-Technical Committee and participated in the compilation by companies such as Sintronic Technology was officially released.


The report shows that face recognition, as the most concerned biometric technology, has won the favor of many companies and research institutions. As of October 2020, there are a total of 10,443 companies across the country whose names, products, brands, and business scope cover "face recognition." The report predicts that the domestic face recognition market size will maintain an average compound growth rate of 23% in the next five years, and the market size will exceed 10 billion yuan by 2024.

The report said that in 2020, the global outbreak of the new coronavirus has caused greater attention to face recognition wearing masks. Commonly used solutions include data enhancement, occlusion recovery, multi-component model fusion, etc., which can be applied to face deployment control, stranger detection, and sensorless traffic. Under the scale of a 300,000-person image library, the accuracy of face recognition with masks can be More than 90%. In addition, a number of domestic technology companies have also achieved certain results in the application of this technology. The "face recognition + AI temperature measurement" access control developed by Jieshun Technology at the beginning of 2020 has become the first batch of "science and technology anti-epidemic" in China. The model.

The report also pointed out that face recognition technology is also facing certain limitations due to similar faces, age, algorithmic bias, diversified scenes, and easier public access to facial images, such as twin recognition problems, security and privacy, etc. Issues such as fairness need to be resolved continuously.

Wechat qr code scanClose
the qr code