(Original title: CFCA security detection service escorts the face recognition live detection algorithm in the financial field)
In recent years, face recognition technology, as a means of identification, has developed rapidly and has been widely used in all walks of life. However, face spoofing attacks from illegal users have brought great security risks to the application of face recognition technology.
In order to solve this problem, face liveness detection technology came into being, which effectively avoids the identity recognition security threat faced by face payment acceptance terminal, face recognition APP, liveness detection algorithm SDK and other applications in the financial field. To this end, the People’s Bank of China issued the “Face Recognition Technology Offline Payment Security Application Specification (Trial)” (hereinfollowing referred to as the “Specification”) in January 2019. General technical requirements, system security requirements, transaction security requirements and other requirements.
In the application of face recognition, the live detection algorithm can verify whether the user is a real living body by combining actions such as blinking, opening the mouth, shaking his head, and nodding, using technologies such as facial key point positioning and face tracking. It can effectively resist common attack methods such as photos, videos, face swaps, masks, occlusions, 3D animations, and screen remakes, thereby helping users identify fraudulent behaviors and protecting their interests. The “Specification” puts forward requirements for both the performance of live detection and performance indicators.
Figure 1. Prosthetic attack performance requirements of live detection algorithm
Figure 2 Requirements for performance indicators of live detection algorithms
China Financial Certification Center (CFCA), as a third-party testing agency, provides security testing services related to face recognition live detection algorithms from multiple perspectives. In addition to covering the relevant requirements of the People’s Bank of China in accordance with the “Specifications”, it will also combine the accumulated experience in project practice. Such as adversarial sample attack, 3D animation, real attack, automated testing, interface security verification and other experience optimization detection services.
In recent years, CFCA has actively deployed in the fields of artificial intelligence, multi-party secure computing, blockchain, and federated learning. It has a number of financial industry authorization qualifications, participates in the formulation of national and financial industry technical standards and testing standards, continues to increase investment and cooperate with All parties in the industry cooperate extensively to provide the financial industry with professional security testing suitable for face recognition live detection algorithms.
CFCA looks forward to working with all parties to jointly promote the standardized and healthy development of face recognition technology in the financial industry.
