First Security Assessments for Face Recognition Systems Begin

Source: People’s Daily Online.

The National Research Center for Information Industry Security has announced that its National Quality Inspection and Testing Center for Speech and Image Recognition Products, known as the National AI Testing Center, has formally launched security assessment activities for face recognition systems.

As artificial intelligence develops rapidly, face recognition has entered public work and life everywhere. Face payment, face check-in and face entry bring convenience, but their own security problems cause worry and unease, and repeated security holes have made practitioners realise that improving the security of face recognition systems themselves is imperative.

The National AI Testing Center is a national-level third-party authoritative testing body for AI under the center, accredited by the Certification and Accreditation Administration in December 2020. It is the first and only national-level speech and image product quality inspection and testing institution in China’s quality inspection system for the AI field.

This is the center’s first assessment campaign focused on the security of face recognition systems. Based on current and in-development national and industry standards, and targeting algorithm security and data security, it will comprehensively assess the security capabilities of face recognition systems against spoofing, injection and adversarial attacks, and in data processing, using presentation attack and injection attack techniques.

As testing and certification, standard-setting and other regulatory work proceed in order, the center will become an important base and technical support institution for quality inspection, technology research and product development in China’s AI industry, consolidating the industry’s quality foundation and leading the industry into a new stage.

← Back to News