Journal of Computer Applications ›› 2021, Vol. 41 ›› Issue (7): 1922-1927.DOI: 10.11772/j.issn.1001-9081.2020081214

Special Issue: 人工智能

• Artificial intelligence • Previous Articles     Next Articles

Generative adversarial network synthesized face detection based on deep alignment network

TANG Guihua, SUN Lei, MAO Xiuqing, DAI Leyu, HU Yongjin   

  1. Information Engineering University, Zhengzhou Hernan 450001, China
  • Received:2020-08-12 Revised:2020-10-23 Online:2021-07-10 Published:2020-11-25
  • Supported by:
    This work is partially supported by the National Key Research and Development Program of China (2016YFB0501900).

基于深度对齐网络的生成对抗网络伪造人脸检测

汤桂花, 孙磊, 毛秀青, 戴乐育, 胡永进   

  1. 信息工程大学, 郑州 450001
  • 通讯作者: 汤桂花
  • 作者简介:汤桂花(1996-),女,四川乐山人,硕士研究生,主要研究方向:图像处理、机器视觉;孙磊(1973-),男,江苏靖江人,教授,博士,主要研究方向:网络空间安全、可信计算;毛秀青(1980-),男,安徽滁州人,副教授,硕士,主要研究方向:信息安全;戴乐育(1990-),男,河南郑州人,讲师,硕士,主要研究方向:神经网络加密、硬件加速;胡永进(1981-),男,山东潍坊人,讲师,硕士,主要研究方向:主动防御、态势感知。
  • 基金资助:
    国家重点研发计划项目(2016YFB0501900)。

Abstract: The existing Generative Adversarial Network (GAN) synthesized face detection method has misjudgment of real faces with angles and occlusion, therefor a GAN-synthesized face detection method based on Deep Alignment Network (DAN) was proposed. Firstly, a facial landmark extraction network was designed based on DAN to extract the locations of facial landmarks of genuinus and synthesized faces. Then, in order to reduce the redundant information and feature dimensionality, each group of landmarks was mapped to the three-dimensional space by using the Principal Component Analysis (PCA) method. Finally, the features were classified by using 5-fold cross-validation of Support Vector Machine and the accuracy was calculated. Experimental results show that the proposed method improves the face dissonance caused by location errors by improving the accuracy of facial landmark location, which reduces the misjudgment rate of real faces. Compared with VGG19, XceptionNet and Dlib-SVM methods, this proposed method has the Area Under Receiver Operating Characteristic curve (AUC) increased by 4.48 to 32.96 percentage points and Average Precision (AP) increased by 4.26 to 33.12 percentage points on frontal faces; and has the AUC increased by 10.56 to 30.75 percentage points and AP increased by 7.42 to 42.45 percentage points on faces with angles and occlusion.

Key words: Generative Adversarial Network (GAN), Deep Alignment Network (DAN), facial landmark, image synthesis, forgery detection

摘要: 针对现有的生成对抗网络(GAN)伪造人脸图像检测方法在有角度及遮挡情况下存在的真实人脸误判问题,提出了一种基于深度对齐网络(DAN)的GAN伪造人脸图像检测方法。首先,基于DAN设计面部关键点提取网络,以提取真伪人脸关键点位置;然后,采用主成分分析(PCA)方法将每一组关键点映射到三维空间,从而减少冗余信息以及降低特征维度;最后,利用支持向量机(SVM)五折交叉验证对特征进行分类,并计算准确率。实验结果表明,该方法通过提高面部关键点定位准确度改善了由于定位误差引起的面部不协调问题,进而降低了真实人脸误判率。与VGG19、XceptionNet和Dlib-SVM方法相比,正脸情况下,该方法的ROC下面积(AUC)值提高了4.48到32.96个百分点,平均精度(AP)提高了4.26到33.12个百分点;有角度及遮挡人脸情况下,该方法的AUC值提高了10.56到30.75个百分点,AP提高了7.42到42.45个百分点。

关键词: 生成对抗网络, 深度对齐网络, 面部关键点, 图像伪造, 伪造检测

CLC Number: