Regularization and attention feature distillation base on light CNN for Hyperspectral face recognition

Regularization and attention feature distillation base on light CNN for Hyperspectral face recognition
复制标题

基于light CNN的正则化和注意力特征蒸馏用于高光谱人脸识别

DOI:
10.1007/s11042-021-10537-4
复制
发表时间:
2021-01
影响因子:
3.6
通讯作者:
Guoyu Lu
Guoyu Lu
中科院分区:
计算机科学4区
文献类型:
--
作者:
Zhihua Xie;Jieyi Niu;Yi Li;Guoyu Lu

文献摘要

参考文献

相似文献

高光谱成像,捕捉一系列光谱波段的判别信息,导致建立一个强大的人脸识别系统。受深度卷积神经网络和迁移学习的成功启发,本文提出了一种基于轻型卷积神经网络(CNN)和迁移学习的端到端高光谱人脸识别模型。为了提高高光谱人脸识别的性能,引入了最大特征映射(MFM)激活函数和微调(结构正则化(L2 SP)或注意力特征蒸馏(AFD))来优化深度网络,它将学习不同波段的精细特征表示。特别地,该方法将正则化和AFD合作纳入可见人脸数据的迁移学习策略中。通过将高光谱图像反馈到预训练的light CNN网络,我们可以设计一个端到端模型,该模型可以利用高光谱人脸图像的泛化能力。最后,整个模型使用相关的标准评估协议在PolyU-HSFD,CMU和UWA高光谱人脸数据集上进行训练和验证。实验结果表明,改进后的Light CNN网络能够很好地表征高光谱人脸特征,L2 SP和AFD相结合的联合训练方法比基于其他深度网络的现有方法具有更好的识别性能。
The hyperspectral imaging, capturing discriminative information across a series of spectrum bands, leads to building a robust face recognition system. Motivated by the success of deep convolutional network and transferring learning, this paper proposed an end-to-end hyperspectral face recognition model based on a light Convolutional Neural Network (CNN) and transfer learning. To boost the performance of hyperspectral face recognition, the Max-Feature-Map (MFM) activation function and fine-tuning (structure regularization(L2SP) or attention feature distillation (AFD)) are introduced to optimize the deep network, which will learn the fine feature representation across different bands. Especially, this method incorporates regularization and AFD cooperation into the transfer learning strategy on the visible face data. By feeding back hyperspectral images to the pretrained light CNN network, we can design an end-to end model that can leverage the generalization ability for hyperspectral face images. Finally, the entire model is trained and verified on the PolyU-HSFD, CMU, and UWA hyperspectral face datasets using the associated standard evaluation protocols. Experimental results demonstrate that the improved Light CNN network can get good representations of hyperspectral face features and the joint training with a combination of L2SP and AFD exhibits better recognition performance than the state-of-the-art methods based on other deep networks.
DOI: 10.1093/pcmedi/pbac012
发表时间: 2022-05-13
影响因子: 5.3
作者:
通讯作者: --
DOI: --
发表时间: 2002
影响因子: 11.4
作者:
F. Kruse
通讯作者: F. Kruse
DOI: --
发表时间: 2019-01
期刊: ArXiv
影响因子: --
作者:
Xingjian Li;Haoyi Xiong;Hanchao Wang;Yuxuan Rao;Liping Liu;Jun Huan
通讯作者: Xingjian Li;Haoyi Xiong;Hanchao Wang;Yuxuan Rao;Liping Liu;Jun Huan
DOI: --
发表时间: 2014-11
期刊: ArXiv
影响因子: --
作者:
Dong Yi;Zhen Lei;Shengcai Liao;S. Li
通讯作者: Dong Yi;Zhen Lei;Shengcai Liao;S. Li
DOI: 10.1007/978-3-319-28501-6_1
发表时间: 2016-02
期刊: 2021 2nd International Conference on Intelligent Engineering and Management (ICIEM)
影响因子: --
作者:
D. Allen
通讯作者: D. Allen