《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (8): 2588-2592.DOI: 10.11772/j.issn.1001-9081.2022071100

• 多媒体计算与计算机仿真 • 上一篇    

基于GhostNet和特征融合的人脸活体检测算法

韩春港1,2, 刘永辉2()   

  1. 1.山东建筑大学 计算机科学与技术学院,济南 250101
    2.浪潮智能终端有限公司 解决方案开发部,济南 250101
  • 收稿日期:2022-07-28 修回日期:2022-11-07 接受日期:2022-11-21 发布日期:2023-01-15 出版日期:2023-08-10
  • 通讯作者: 韩春港
  • 作者简介:刘永辉(1975—),男,山东济南人,高级工程师,硕士,主要研究方向:音视频处理、机器视觉、大数据分析。

Face liveness detection algorithm based on GhostNet and feature fusion

Chungang HAN1,2, Yonghui LIU2()   

  1. 1.School of Computer Science and Technology,Shandong Jianzhu University,Jinan Shandong 250101,China
    2.Solution Development Department,Inspur Intelligent Terminal Company Limited,Jinan Shandong 250101,China
  • Received:2022-07-28 Revised:2022-11-07 Accepted:2022-11-21 Online:2023-01-15 Published:2023-08-10
  • Contact: Chungang HAN
  • About author:LIU Yonghui, born in 1975, M. S., senior engineer. His research interests include audio and video processing, machine vision, big data analysis. 837164564@qq.com

摘要:

人脸识别技术的广泛应用在为用户带来方便的同时,也带来了人脸欺骗和展示攻击等问题。针对经常出现的展示攻击和打印攻击问题,提出了一种基于GhostNet和特征融合的人脸活体检测算法。首先,将GhostNet模型的特征提取过程分为三种不同的阶段,即低等特征、中等特征和高等特征;然后,分别输出每个阶段的特征图信息;最后,将具有不同语义信息的特征图送入特征融合模块进行自适应加权融合,以获得更加具有辨别性的特征映射。在NUAA和CelebA-Spoof两个公开数据集上进行实验,实验结果表明所提算法的准确率分别为99.97%和93.41%,相较于GhostNet模型直接进行训练的算法分别提高了8.00和9.20个百分点。与异构内核的卷积神经网络(HK-CNN)、轻量级卷积神经网络FeatherNet、基于分块的多流网络FaceBagNet等算法相比,所提算法在NUAA和CelebA-Spoof数据集上表现出更好的性能;并且,由于GhostNet是一种轻量化的网络模型,所提算法在CelebA-Spoof数据集上对单张图像进行推理的时间仅需3.6 ms

关键词: 人脸识别, 活体检测, GhostNet, 自适应加权融合, CelebA-Spoof数据集

Abstract:

The wide application of face recognition technology not only brings convenience to users, but also brings problems such as face spoofing and presentation attacks. Aiming at the frequent presentation attacks and print attacks, a face liveness detection algorithm based on GhostNet and feature fusion was proposed. Firstly, the feature extraction process of GhostNet model was divided into three different stages, namely, low-level feature, medium-level feature and high-level feature. Then, the feature map information of each stage was output respectively. Finally, the feature maps with different semantic information were sent into the feature fusion module for adaptive weighted fusion, so as to obtain more discriminative feature mapping. Experiments were conducted on public datasets NUAA and CelebA-Spoof. The results show that the accuracy of the proposed algorithm is 99.97% and 93.41% respectively, which is increased by 8.00 and 9.20 percentage points respectively compared with the algorithm of direct training of GhostNet model. Compared with Heterogeneous Kernel-Convolutional Neural Network (HK-CNN), lightweight convolutional neural network FeatherNet, block based multi-stream network FaceBageNet and other algorithms, the proposed algorithm shows better performance on NUAA and CelebA-Spoof datasets. And, as GhostNet is a lightweight network model, the proposed algorithm only takes 3.6 ms on single image inference on CelebA-Spoof dataset.

Key words: face recognition, live detection, GhostNet, adaptive weighted fusion, CelebA-Spoof dataset

中图分类号: