本文已被:浏览 35次 下载 444次
Received:May 08, 2024 Revised:May 29, 2024
Received:May 08, 2024 Revised:May 29, 2024
中文摘要: 本文提出了一种将多尺度频率特征和生成对抗网络(GAN)训练的深度图特征融合的多分支网络. 具体地, 高频特征中的边缘纹理信息有利于捕捉摩尔纹. 低频特征对色彩失真更为敏感. 作为辅助信息, 深度图在视觉层面上比 RGB 图像更具辨别力. 有监督多视图对比学习的应用进一步增强了多视图特征的学习. 此外, 还提出了两阶段双线性特征融合方法, 以融合来自不同视图的多分支特征. 为了评估该模型, 我们在4个广泛使用的公共数据集(CASIA-FASD、Replay-Attack、MSU-MFSD 和 OULU-NPU)上进行了消融实验, 特征融合对比实验, 单一数据集实验和跨数据集实验. 跨数据集实验结果表明, 本文模型在4种测试协议上的平均HTER比只使用RGB图转换为深度图(DFA)的方法好5% (20.3%减至15.0%).
Abstract:In this study, a multi-branch network that integrates multi-scale frequency features and depth map features trained by generative adversarial network (GAN) is proposed. Specifically, edge texture information in high-frequency features is beneficial to capturing moire patterns. Low-frequency features are more sensitive to color distortion. Depth maps are more discriminative than RGB images from the visual level as auxiliary information. Supervised multi-view contrastive learning is employed to further enhance multi-view feature learning. Moreover, a two-stage bilinear feature fusion method is proposed to effectively integrate multi-branch features from different views. To evaluate the model, ablation experiments, feature fusion comparison experiments, intra-set experiments and inter-set experiments are conducted on four widely used public datasets, namely CASIA-FASD, Replay-Attack, MSU-MFSD, and OULU-NPU. The experiment result shows that the average HTER of the proposed model on the four tested protocols is 5% (20.3% to 15.0%) better than the DFA method in the inter-set evaluation.
keywords: face anti-spoofing (FAS) contrastive learning feature fusion generative adversarial network (GAN) deep learning
文章编号: 中图分类号: 文献标志码:
基金项目:
引用文本:
孙文赟,李进,金忠.有监督多视图对比学习和两阶段双线性特征融合的人脸活体检测.计算机系统应用,2024,33(11):131-141
SUN Wen-Yun,LI Jin,JIN Zhong.Face Anti-spoofing Based on Supervised Multi-view Contrastive Learning and Two-stage Bilinear Feature Fusion.COMPUTER SYSTEMS APPLICATIONS,2024,33(11):131-141
孙文赟,李进,金忠.有监督多视图对比学习和两阶段双线性特征融合的人脸活体检测.计算机系统应用,2024,33(11):131-141
SUN Wen-Yun,LI Jin,JIN Zhong.Face Anti-spoofing Based on Supervised Multi-view Contrastive Learning and Two-stage Bilinear Feature Fusion.COMPUTER SYSTEMS APPLICATIONS,2024,33(11):131-141