面向不同传感器与复杂场景的人脸识别系统防伪方法综述

人脸识别技术作为一项重要的生物特征识别技术，在人们的日常生活中得到广泛应用。尽管人脸识别技术已取得明显进展，但当前的人脸识别系统仍容易受到非法用户的恶意攻击，因此人脸防伪技术成为人脸识别过程中必不可少的一个环节。在简述人脸防伪概念及常见攻击类型的基础上，分析人脸防伪方法的主要建模思路，从面向不同传感器与面向复杂场景的人脸识别系统防伪方法切入，分类阐述不同人脸防伪方法的基本原理及发展脉络。总结公开人脸防伪数据库，分析比较代表性防伪方法及其性能表现，并对人脸防伪问题的未来发展趋势进行展望。

Overview of Anti-Spoofing Methods of Face Recognition Systems for Different Sensors and Complex Scenes

As an important biometric recognition technology, face recognition has been widely used in daily life.Although significant progress has been made in this field, the existing face recognition systems are vulnerable to malicious attacks from illegal users.Therefore, face anti-spoofing technology has become an indispensable step in the procedure of face recognition.This paper introduces the basic concept of face anti-spoofing and the common attack types, and analyzes the main modeling ideas of face anti-spoofing.Then the paper focuses on the anti-spoofing methods in face recognition systems for different sensors and complex scenarios, describing the principles and the development history of different face anti-spoofing methods.In addition, the paper lists public face anti-spoofing databases, and on this basis compares the performance of typical anti-spoofing methods on these databases.Finally, the paper discusses the future development trend of face anti-spoofing.