基于移动端的高效人脸识别算法
|
摘要
随着深度学习的快速发展,深度卷积神经网络已广泛用于人脸识别领域,但深度模型推理的高计算复杂度和大量的参数阻碍人脸识别系统的实际部署。因此提出两种改进算法,一是基于深度特征蒸馏的人脸识别方法,该方法以预训练的大网络为教师网络,指导小网络训练,将知识迁移得到轻量级的学生网络,提出新的人脸损失函数。二是对教师网络和学生网络分别添加SENet网络结构,使人脸识别准确率有很大的提升。通过实验验证,学生网络模型大小压缩至3.22M,相比于教师模型压缩4倍,在LFW数据集上准确率达到99.66%。
With the rapid development of deep learning, deep convolutional neural networks have been widely used in the field of face recognition, but the high computational complexity and a large number of parameters of deep model inference hinder the actual deployment of face recogni?tion systems. Therefore, this paper proposes two improved algorithms. One is the face recognition method based on deep feature distillation.This method uses pre-trained large network as the teacher network to guide small network training, and transfer knowledge to a lightweight student network. Face loss function. The second is to add the SENet network structure to the teacher network and the student network, so that the face recognition accuracy rate is greatly improved. Through experimental verification, the student network model size is compressed to 3.22 M, which is 4 times smaller than the teacher model and 99.66% in the LFW data set.
With the rapid development of deep learning, deep convolutional neural networks have been widely used in the field of face recognition, but the high computational complexity and a large number of parameters of deep model inference hinder the actual deployment of face recogni?tion systems. Therefore, this paper proposes two improved algorithms. One is the face recognition method based on deep feature distillation.This method uses pre-trained large network as the teacher network to guide small network training, and transfer knowledge to a lightweight student network. Face loss function. The second is to add the SENet network structure to the teacher network and the student network, so that the face recognition accuracy rate is greatly improved. Through experimental verification, the student network model size is compressed to 3.22 M, which is 4 times smaller than the teacher model and 99.66% in the LFW data set.
引文
[1]Simonyan K,Zisserman A. Very Deep Convolutional Networks for Large-Scale Image Recognition[J]. arXiv preprint arXiv:1409.1556,2014.
[2]Szegedy C,Liu W,Jia Y,et al. Going Deeper with Convolutions[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2015:1-9.
[3]He K,Zhang X,Ren S,et al. Deep Residual Learning for Image Recognition[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2016:770-778.
[4]Hinton G,Vinyals O,Dean J. Distilling the Knowledge in a Neural Network[J]. arXiv Preprint arXiv:1503.02531,2015.
[5]Romero A,Ballas N,Kahou S E,et al. FitNets:Hints for Thin Deep Nets[J]. arXiv Preprint arXiv:1412.6550,2014.
[6]Hu J,Shen L,Sun G. Squeeze-and-Excitation Networks[J]. arXiv preprint arXiv:1709.01507,2017,7.
[7]Huang,G.B.,Ramesh,M.,Berg,T.,et al. Labeled Faces in the Wild:a Database for Studying Face Recognition in Unconstrained Environments,2007.
[8]Yaniv Taigman,Ming Yang,Marcaurelio Ranzato,Lior Wolf. DeepFace:Closing the Gap to Human-Level Performance in Face Verification. Computer Vision and Pattern Recognition,2014.
[9]Florian Schroff,Dmitry Kalenichenko,James Philbin. FaceNet:A Unified Embedding for Face Recognition and Clustering. Computer Vision and Pattern Recognition,2015.
[10]W. Liu,Y. Wen,Z.Yu,M. Yang. Large-Margin Softmax Loss for Convolutional Neural Networks. In ICML,2016.2,3,7,8.
[11]W.Liu,Y. Wen,Z. Yu,M. Li,B. Raj,L. Song. Sphereface:Deep Hypersphere Embedding for Face Recognition. arXiv preprint arXiv:1704.08063,2017.2,8.
[12]F. Wang,W. Liu,H. Liu,J. Cheng. Additive Margin Softmax for Face Verification. In arXiv:1801.05599,2018.1,2,3,4,9.
[13]Deng J,Guo J,Zafeiriou S. ArcFace:Additive Angular Margin Loss for Deep Face Recognition[J]. arXiv preprint arXiv:1801.07698,2018.
[14]Iandola F N,Han S,Moskewicz M W,et al. Squeezenet:AlexNet-Level Accuracy with 50x Fewer Parameters and<0.5 MB Model Size[J]. arXiv preprint arXiv:1602.07360,2016.
[15]Krizhevsky A,Sutskever I,Hinton G E. ImageNet Classification with Deep Convolutional Neural Networks[C]. Advances in Neural Information Processing Systems. 2012:1097-1105.
[16]Howard A G,Zhu M,Chen B,et al. MobileNets:Efficient Convolutional Neural Networks for Mobile Vision Applications[J]. arXiv Preprint arXiv:1704.04861,2017.
[17]Zhang X,Zhou X,Lin M,et al. ShuffleNet:An Extremely Efficient Convolutional Neural Network for Mobile Devices. arXiv Preprint arXiv:1707.01083.
[18]Sandler M,Howard A,Zhu M,et al. Inverted Residuals and Linear Bottlenecks:Mobile Networks for Classification,Detection and Segmentation[J]. arXiv preprint arXiv:1801.04381,2018.
[19]Howard A G,Zhu M,Chen B,et al. Mobilenets:Efficient Convolutional Neural Networks for Mobile Vision Applications[J]. arXiv Preprint arXiv:1704.04861,2017.
[20]Luo P,Zhu Z,Liu Z,et al. Face Model Compression by Distilling Knowledge from Neurons[C]. AAAI. 2016:3560-3566.
[21]Sun Y,Chen Y,Wang X,et al. Deep Learning Face Representation by Joint Identification-Verification[C]. Advances in Neural Information Processing Systems. 2014:1988-1996.
[22]Kemelmacher-Shlizerman I,Seitz S M,Miller D,et al. The MegaFace Benchmark:1 Million Faces for Recognition at Scale[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016:4873-4882.
[23]Zhang K,Zhang Z,Li Z,et al. Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks[J]. IEEE Signal Processing Letters,2016,23(10):1499-1503.
[2]Szegedy C,Liu W,Jia Y,et al. Going Deeper with Convolutions[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2015:1-9.
[3]He K,Zhang X,Ren S,et al. Deep Residual Learning for Image Recognition[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2016:770-778.
[4]Hinton G,Vinyals O,Dean J. Distilling the Knowledge in a Neural Network[J]. arXiv Preprint arXiv:1503.02531,2015.
[5]Romero A,Ballas N,Kahou S E,et al. FitNets:Hints for Thin Deep Nets[J]. arXiv Preprint arXiv:1412.6550,2014.
[6]Hu J,Shen L,Sun G. Squeeze-and-Excitation Networks[J]. arXiv preprint arXiv:1709.01507,2017,7.
[7]Huang,G.B.,Ramesh,M.,Berg,T.,et al. Labeled Faces in the Wild:a Database for Studying Face Recognition in Unconstrained Environments,2007.
[8]Yaniv Taigman,Ming Yang,Marcaurelio Ranzato,Lior Wolf. DeepFace:Closing the Gap to Human-Level Performance in Face Verification. Computer Vision and Pattern Recognition,2014.
[9]Florian Schroff,Dmitry Kalenichenko,James Philbin. FaceNet:A Unified Embedding for Face Recognition and Clustering. Computer Vision and Pattern Recognition,2015.
[10]W. Liu,Y. Wen,Z.Yu,M. Yang. Large-Margin Softmax Loss for Convolutional Neural Networks. In ICML,2016.2,3,7,8.
[11]W.Liu,Y. Wen,Z. Yu,M. Li,B. Raj,L. Song. Sphereface:Deep Hypersphere Embedding for Face Recognition. arXiv preprint arXiv:1704.08063,2017.2,8.
[12]F. Wang,W. Liu,H. Liu,J. Cheng. Additive Margin Softmax for Face Verification. In arXiv:1801.05599,2018.1,2,3,4,9.
[13]Deng J,Guo J,Zafeiriou S. ArcFace:Additive Angular Margin Loss for Deep Face Recognition[J]. arXiv preprint arXiv:1801.07698,2018.
[14]Iandola F N,Han S,Moskewicz M W,et al. Squeezenet:AlexNet-Level Accuracy with 50x Fewer Parameters and<0.5 MB Model Size[J]. arXiv preprint arXiv:1602.07360,2016.
[15]Krizhevsky A,Sutskever I,Hinton G E. ImageNet Classification with Deep Convolutional Neural Networks[C]. Advances in Neural Information Processing Systems. 2012:1097-1105.
[16]Howard A G,Zhu M,Chen B,et al. MobileNets:Efficient Convolutional Neural Networks for Mobile Vision Applications[J]. arXiv Preprint arXiv:1704.04861,2017.
[17]Zhang X,Zhou X,Lin M,et al. ShuffleNet:An Extremely Efficient Convolutional Neural Network for Mobile Devices. arXiv Preprint arXiv:1707.01083.
[18]Sandler M,Howard A,Zhu M,et al. Inverted Residuals and Linear Bottlenecks:Mobile Networks for Classification,Detection and Segmentation[J]. arXiv preprint arXiv:1801.04381,2018.
[19]Howard A G,Zhu M,Chen B,et al. Mobilenets:Efficient Convolutional Neural Networks for Mobile Vision Applications[J]. arXiv Preprint arXiv:1704.04861,2017.
[20]Luo P,Zhu Z,Liu Z,et al. Face Model Compression by Distilling Knowledge from Neurons[C]. AAAI. 2016:3560-3566.
[21]Sun Y,Chen Y,Wang X,et al. Deep Learning Face Representation by Joint Identification-Verification[C]. Advances in Neural Information Processing Systems. 2014:1988-1996.
[22]Kemelmacher-Shlizerman I,Seitz S M,Miller D,et al. The MegaFace Benchmark:1 Million Faces for Recognition at Scale[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016:4873-4882.
[23]Zhang K,Zhang Z,Li Z,et al. Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks[J]. IEEE Signal Processing Letters,2016,23(10):1499-1503.