卷积神经网络结构优化综述

林景栋; 吴欣怡; 柴毅; 尹宏鹏

doi:10.16383/j.aas.c180275

卷积神经网络结构优化综述

doi: 10.16383/j.aas.c180275

林景栋^1,,
吴欣怡^1,,
柴毅^1,,
尹宏鹏^1, ,

1.
重庆大学自动化学院重庆 400044

基金项目:

国家自然科学基金 61633005

国家自然科学基金 61773080

中央高校基本科研业务费专项资金 2019CDYGZD001

重庆市基础科学与研究技术专项 cstc2015jcyjB0569

重庆大学科研后备拔尖人才 cqu2018CDHB1B04

重庆市重点科技专项子项 cstc2015shms-ztzx30001

详细信息

作者简介:
林景栋  重庆大学自动化学院副教授. 2002年获得重庆大学博士学位.主要研究方向为工业自动化生产线设计, 智能家居控制系统的设计.E-mail: linzhanding@163.com

吴欣怡  重庆大学自动化学院硕士研究生. 2016年获得重庆大学学士学位.主要研究方向为深度学习, 计算机视觉. E-mail: wuxinyi12358@gmail.com

柴毅  重庆大学自动化学院教授.2001年获得重庆大学博士学位.主要研究方向为信息处理, 融合与控制, 计算机网络与系统控制. E-mail:chaiyi@cqu.edu.cn

通讯作者:
尹宏鹏重庆大学自动化学院教授.2009年获得重庆大学博士学位.主要研究方向为模式识别与智能系统.本文通信作者.E-mail:yinhongpeng@gmail.com

计量
- 文章访问数: 8448
- HTML全文浏览量: 3390
- PDF下载量: 2410
- 被引次数: 0
出版历程
- 收稿日期: 2018-05-03
- 录用日期: 2018-11-05
- 刊出日期: 2020-01-21

Structure Optimization of Convolutional Neural Networks: A Survey

LIN Jing-Dong^1
,,
WU Xin-Yi^1
,,
CHAI Yi^1
,,
YIN Hong-Peng^{1
, ,}

1.
College of Automation, Chongqing University, Chongqing 400044

Funds:

National Natural Science Foundation of China 61633005

National Natural Science Foundation of China 61773080

Fundamental Research Funds for the Central Universities 2019CDYGZD001

Chongqing Nature Science Foundation of Fundamental Science and Frontier Technologies cstc2015jcyjB0569

Scientiflc Reserved Talents of Chongqing University cqu2018CDHB1B04

Chongqing Nature Science Foundation of Scientiflc Key Program cstc2015shms-ztzx30001

More Information

Author Bio:
LIN Jing-Dong   Associate professor at the College of Automation, Chongqing University. He received his Ph. D. degree from Chongqing University in 2002. His research interest covers industrial automation line design, and smart home control system design

WU Xin-Yi   Master student at the College of Automation, Chongqing University. He received his bachelor degree from Chongqing University in 2016. His research interest covers deep learning and computer vision

CHAI Yi   Professor at the College of Automation, Chongqing University. He received his Ph. D. degree from Chongqing University in 2001. His research interest covers information processing, integration and control, and computer network and system control

Corresponding author: YIN Hong-Peng Professor at the College of Automation, Chongqing University. He received his Ph. D. degree from Chongqing University in 2009. His research interest covers pattern recognition, image processing, and computer vision. Corresponding author of this paper

摘要

摘要: 近年来, 卷积神经网络(Convolutional neural network, CNNs)在计算机视觉、自然语言处理、语音识别等领域取得了突飞猛进的发展, 其强大的特征学习能力引起了国内外专家学者广泛关注.然而, 由于深度卷积神经网络普遍规模庞大、计算度复杂, 限制了其在实时要求高和资源受限环境下的应用.对卷积神经网络的结构进行优化以压缩并加速现有网络有助于深度学习在更大范围的推广应用, 目前已成为深度学习社区的一个研究热点.本文整理了卷积神经网络结构优化技术的发展历史、研究现状以及典型方法, 将这些工作归纳为网络剪枝与稀疏化、张量分解、知识迁移和精细模块设计4个方面并进行了较为全面的探讨.最后, 本文对当前研究的热点与难点作了分析和总结, 并对网络结构优化领域未来的发展方向和应用前景进行了展望.
- 卷积神经网络 /
- 结构优化 /
- 网络剪枝 /
- 张量分解 /
- 知识迁移
Abstract: Recently convolutional neural networks (CNNs) have made great progress in computer vision, natural language processing and speech recognition, which attracts wide attention for their powerful ability of feature learning. However, deep convolutional neural networks usually have large capacity and high computational complexity, hindering their applications in real-time and source-constrained areas. Thus, optimizing the structure of deep model will contribute to rapid deployment of such networks, which has been a hot topic of deep learning community. In this paper, we provide a comprehensive survey of history progress, recent advances and typical approaches in network structure optimization. These approaches are mainly categorized into four schemes, which are pruning & sparsification, tensor factorization, knowledge transferring and compacting module designing. Finally, the remaining problems and potential trend in this topic are concluded and discussed.
- Convolutional neural networks (CNNs) /
- structure optimization /
- network pruning /
- tensor factorization /
- knowledge transferring
Recommended by Associate Editor HE Wei
注释:

1) 本文责任编委贺威

HTML全文

图 1 LeNet-5网络结构^[6]

Fig. 1 Structure of LeNet-5^[6]

下载: 全尺寸图片幻灯片

图 2 四种剪枝粒度方式^[26]

Fig. 2 Four pruning granularities^[26]

下载: 全尺寸图片幻灯片

图 3 张量分解过程

Fig. 3 Process of tensor factorization

下载: 全尺寸图片幻灯片

图 4 知识迁移过程

Fig. 4 Process of knowledge transfer

下载: 全尺寸图片幻灯片

图 5 Inception-v1结构^[4]

Fig. 5 Inception-v1 module^[4]

下载: 全尺寸图片幻灯片

图 6 卷积核分解示意图^[57]

Fig. 6 Process of convolutional filter factorization^[57]

下载: 全尺寸图片幻灯片

图 7 Xception模块^[59]

Fig. 7 Xception module^[59]

下载: 全尺寸图片幻灯片

图 8 线性卷积结构与多层感知机卷积结构^[55]

Fig. 8 Linear convolutional structure and Mlpconv structure^[55]

下载: 全尺寸图片幻灯片

图 9 残差模块^[5]

Fig. 9 Residual module^[5]

下载: 全尺寸图片幻灯片

表 1 经典卷积神经网络的性能及相关参数

Table 1 Classic convolutional neural networks and corresponding parameters

年份	网络名称	网络层数	卷积层数量	参数数量卷积层全连接层		乘加操作数（MACs) 卷积层全连接层		Top-5错误率（％)
2012	AlexNet^[1]	8	5	2.3M	58.6M	666 M	58.6M	16.4
2014	Overfeat^[2]	8	5	16 M	130M	2.67G	124M	14.2
2014	VGGNet-16^[3]	16	13	14.7M	124M	15.3 G	130M	7.4
2015	GoogLeNet^[4]	22	21	6M	1M	1.43 G	1M	6.7
2016	ResNet-50^[5]	50	49	23.5M	2M	3.86 G	2M	3.6

下载: 导出CSV

表 2 网络剪枝对不同网络的压缩效果

Table 2 Comparison of different pruned networks

所用方法	选用网络	初始错误率	剪枝后错误率	初始参数量	剪枝后参数量	压缩率
[28]	AlexNet	19.73%	19.70%	61 M	6.7M	6×
[29]	CaffeNet	42.16%	44.4%	61 M	21.3M	3×
[30]	LeNet-5	0.91%	0.91 %	431 K	4.0 K	108×
[33]	VGGNet-16	6.75%	6.6%	150 M	5.4M	28×
[34]	ResNet-50	8.86%	11.7%	25.56 M	8.66M	3×

下载: 导出CSV

参考文献(79)

[1]	Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks. In: Proceedings of the 25th International Conference on Neural Information Processing Systems. Lake Tahoe, Nevada, USA: Curran Associates Inc., 2012. 1097-1105
[2]	Zeiler M D, Fergus R. Visualizing and understanding convolutional networks. In: Proceedings of 13th European Conference on Computer Vision. Zurich, Switzerland: Springer, 2014. 818-833 http://cn.bing.com/academic/profile?id=2e04eadd73b7358f1ea104aef2c94bd4&encoded=0&v=paper_preview&mkt=zh-cn
[3]	Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. arXiv: 1409.1556, 2014. http://cn.bing.com/academic/profile?id=9a83dddfc646cd21a3e38737d303a369&encoded=0&v=paper_preview&mkt=zh-cn
[4]	Szegedy C, Liu W, Jia Y Q, Sermanet P, Reed S, Anguelov D, et al. Going deeper with convolutions. In: Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA: IEEE, 2015. 1-9 http://cn.bing.com/academic/profile?id=7d4011aa0a4959f0c5e4af61acc12466&encoded=0&v=paper_preview&mkt=zh-cn
[5]	He K M, Zhang X Y, Ren S Q, Sun J. Deep residual learning for image recognition. In: Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016. 770-778 http://cn.bing.com/academic/profile?id=d3fa279e4a35560a5429ba8f84dff15e&encoded=0&v=paper_preview&mkt=zh-cn
[6]	LeCun Y, Bottou L, Bengio Y, Haffner P. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 1998, 86(11): 2278-2324 doi: 10.1109/5.726791
[7]	He K M, Sun J. Convolutional neural networks at constrained time cost. In: Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, USA: IEEE, 2015. 5353-5360
[8]	Le Cun Y, Denker J S, Solla S A. Optimal brain damage. In: Proceedings of the 2nd International Conference on Neural Information Processing Systems. Denver, Colorado, USA: MIT Press, 1989. 598-605
[9]	Hassibi B, Stork D G, Wolff G, Watanabe T. Optimal brain surgeon: extensions and performance comparisons. In: Proceedings of the 6th International Conference on Neural Information Processing Systems. Denver, Colorado, USA: Morgan Kaufmann Publishers Inc., 1993. 263-270
[10]	Cheng Y, Wang D, Zhou P, Zhang T. A survey of model compression and acceleration for deep neural networks. arXiv: 1710.09282, 2017.
[11]	Cheng J, Wang P S, Li G, Hu Q H, Lu H Q. Recent advances in efficient computation of deep convolutional neural networks. Frontiers of Information Technology & Electronic Engineering, 2018, 19(1): 64-77 http://d.old.wanfangdata.com.cn/Periodical/zjdxxbc-e201801008
[12]	雷杰, 高鑫, 宋杰, 王兴路, 宋明黎.深度网络模型压缩综述.软件学报, 2018, 29(2): 251-266 http://d.old.wanfangdata.com.cn/Periodical/rjxb201802002 Lei Jie, Gao Xin, Song Jie, Wang Xing-Lu, Song Ming-Li. Survey of deep neural network model compression. Journal of Software, 2018, 29(2): 251-266 http://d.old.wanfangdata.com.cn/Periodical/rjxb201802002
[13]	Hu H Y, Peng R, Tai Y W, Tang C K. Network trimming: a data-driven neuron pruning approach towards efficient deep architectures. arXiv: 1607.03250, 2016.
[14]	Cheng Y, Wang D, Zhou P, Zhang T. Model compression and acceleration for deep neural networks: the principles, progress, and challenges. IEEE Signal Processing Magazine, 2018, 35(1): 126-136 http://cn.bing.com/academic/profile?id=c41edb9c79f4cae56125bbfc508801d3&encoded=0&v=paper_preview&mkt=zh-cn
[15]	Gong Y C, Liu L, Yang M, Bourdev L. Compressing deep convolutional networks using vector quantization. arXiv: 1412.6115, 2014. http://cn.bing.com/academic/profile?id=fb878b7aaad93122079eeaf80c4b058f&encoded=0&v=paper_preview&mkt=zh-cn
[16]	Reed R. Pruning algorithms-a survey. IEEE Transactions on Neural Networks, 1993, 4(5): 740-747 doi: 10.1109/72.248452
[17]	Collins M D, Kohli P. Memory bounded deep convolutional networks. arXiv: 1412.1442, 2014.
[18]	Jin X J, Yuan X T, Feng J S, Yan S C. Training skinny deep neural networks with iterative hard thresholding methods. arXiv: 1607.05423, 2016.
[19]	Zhou H, Alvarez J M, Porikli F. Less is more: towards compact CNNs. In: Proceedings of the 14th European Conference on Computer Vision. Amsterdam, The Netherlands: Springer, 2016. 662-677
[20]	Wen W, Wu C P, Wang Y D, Chen Y R, Li H. Learning structured sparsity in deep neural networks. In: Proceedings of the 30th Conference on Neural Information Processing Systems. Barcelona, Spain: MIT Press, 2016. 2074-2082
[21]	Lebedev V, Lempitsky V. Fast convnets using group-wise brain damage. In: Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, USA: IEEE, 2016. 2554-2564
[22]	Louizos C, Welling M, Kingma D P. Learning sparse neural networks through L₀regularization. arXiv: 1712.01312, 2017.
[23]	Hinton G E, Srivastava N, Krizhevsky A, Sutskever I, Salakhutdinov R R. Improving neural networks by preventing co-adaptation of feature detectors. arXiv: 1207.0580, 2012.
[24]	Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R. Dropout: a simple way to prevent neural networks from overfitting. Journal of Machine Learning Research, 2014, 15(1): 1929-1958 http://d.old.wanfangdata.com.cn/Periodical/kzyjc200606005
[25]	Li Z, Gong B Q, Yang T B. Improved dropout for shallow and deep learning. In: Proceedings of the 30th Conference on Neural Information Processing Systems. Barcelona, Spain: MIT Press, 2016. 2523-2531
[26]	Anwar S, Sung W. Coarse pruning of convolutional neural networks with random masks. In: Proceedings of 2017 International Conference on Learning Representations. Toulon, France: 2017. 134-145
[27]	Hanson S J, Pratt L Y. Comparing biases for minimal network construction with back-propagation. In: Proceedings of the 1st International Conference on Neural Information Processing Systems. Denver, Colorado, USA: MIT Press, 1988. 177-185
[28]	Han S, Mao H Z, Dally W J. Deep compression: compressing deep neural networks with pruning, trained quantization and huffman coding. arXiv: 1510.00149, 2015. http://cn.bing.com/academic/profile?id=9bf6fa99e4da3298640c577b462462d5&encoded=0&v=paper_preview&mkt=zh-cn
[29]	Srinivas S, Babu R V. Data-free parameter pruning for deep neural networks. arXiv: 1507.06149, 2015.
[30]	Guo Y W, Yao A B, Chen Y R. Dynamic network surgery for efficient DNNs. In: Proceedings of the 30th Conference on Neural Information Processing Systems. Barcelona, Spain: MIT Press, 2016. 1379-1387
[31]	Liu X Y, Pool J, Han S, Dally W J. Efficient sparse-winograd convolutional neural networks. In: Proceedings of 2017 International Conference on Learning Representation. France: 2017.
[32]	He Y H, Zhang X Y, Sun J. Channel pruning for accelerating very deep neural networks. In: Proceedings of 2017 IEEE International Conference on Computer Vision. Venice, Italy: IEEE, 2017. 1398-1406
[33]	Li H, Kadav A, Durdanovic I, Samet H, Graf H P. Pruning filters for efficient convNets. arXiv: 1608.08710, 2016.
[34]	Luo J H, Wu J X, Lin W Y. Thinet: a filter level pruning method for deep neural network compression. In: Proceedings of 2017 IEEE International Conference on Computer Vision. Venice, Italy: IEEE, 2017. 5068-5076
[35]	Denil M, Shakibi B, Dinh L, Ranzato M, de Freitas N. Predicting parameters in deep learning. In: Proceedings of the 26th International Conference on Neural Information Processing Systems. Lake Tahoe, Nevada, USA: Curran Associates Inc., 2013. 2148-2156
[36]	Rigamonti R, Sironi A, Lepetit V, Fua P. Learning separable filters. In: Proceedings of 2013 IEEE Conference on Computer Vision and Pattern Recognition. Portland, OR, USA: IEEE, 2013. 2754-2761
[37]	Jaderberg M, Vedaldi A, Zisserman A. Speeding up convolutional neural networks with low rank expansions. arXiv: 1405.3866, 2014. http://cn.bing.com/academic/profile?id=2c2b54ee2cf492b32a9efa47b48a5cfc&encoded=0&v=paper_preview&mkt=zh-cn
[38]	Denton E, Zaremba W, Bruna J, LeCun Y, Fergus R. Exploiting linear structure within convolutional networks for efficient evaluation. In: Proceedings of the 27th International Conference on Neural Information Processing Systems. Montreal, Canada: MIT Press, 2014. 1269-1277
[39]	Lebedev V, Ganin Y, Rakhuba M, Oseledets I, Lempitsky V. Speeding-up convolutional neural networks using fine-tuned CP-decomposition. arXiv: 1412.6553, 2014.
[40]	Tai C, Xiao T, Zhang Y, Wang X G, E W N. Convolutional neural networks with low-rank regularization. arXiv: 1511.06067, 2015.
[41]	Zhang X Y, Zou J H, Ming X, He K M, Sun J. Efficient and accurate approximations of nonlinear convolutional networks. In: Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, USA: IEEE, 2015. 1984-1992
[42]	Kim Y D, Park E, Yoo S, Choi T, Yang L, Shin D. Compression of deep convolutional neural networks for fast and low power mobile applications. arXiv: 1511.06530, 2015. http://cn.bing.com/academic/profile?id=281aebb382c2e11ab8d73baaafadfbe5&encoded=0&v=paper_preview&mkt=zh-cn
[43]	Wang Y H, Xu C, Xu C, Tao D C. Beyond filters: compact feature map for portable deep model. In: Proceedings of the 34th International Conference on Machine Learning. Sydney, Australia: JMLR.org, 2017. 3703-3711
[44]	Astrid M, Lee S I. CP-decomposition with tensor power method for convolutional neural networks compression. In: Proceedings of 2017 IEEE International Conference on Big Data and Smart Computing. Jeju, South Korea: IEEE, 2017. 115-118
[45]	Bucilu\v{a} C, Caruana R, Niculescu-Mizil A. Model compression. In: Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Philadelphia, USA: ACM, 2006. 535-541
[46]	Ba J, Caruana R. Do deep nets really need to be deep? In: Proceedings of Advances in Neural Information Processing Systems. Montreal, Quebec, Canada: MIT Press, 2014. 2654-2662
[47]	Hinton G, Vinyals O, Dean J. Distilling the knowledge in a neural network. arXiv: 1503.02531, 2015.
[48]	Romero A, Ballas N, Kahou S E, Chassang A, Gatta C, Bengio Y. Fitnets: hints for thin deep nets. arXiv: 1412.6550, 2014.
[49]	Luo P, Zhu Z Y, Liu Z W, Wang X G, Tang X O. Face model compression by distilling knowledge from neurons. In: Proceedings of the 30th AAAI Conference on Artificial Intelligence. Phoenix, Arizona, USA: AAAI, 2016. 3560-3566
[50]	Chen T Q, Goodfellow I, Shlens J. Net2Net: accelerating learning via knowledge transfer. arXiv: 1511.05641, 2015.
[51]	Zagoruyko S, Komodakis N. Paying more attention to attention: improving the performance of convolutional neural networks via attention transfer. In: Proceedings of 2017 International Conference on Learning Representations. France: 2017.
[52]	Theis L, Korshunova I, Tejani A, Huszár F. Faster gaze prediction with dense networks and Fisher pruning. arXiv: 1801.05787, 2018.
[53]	Yim J, Joo D, Bae J, Kim J. A gift from knowledge distillation: fast optimization, network minimization and transfer learning. In: Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI, USA: IEEE, 2017.
[54]	Chen G B, Choi W, Yu X, Han T, Chandraker M. Learning efficient object detection models with knowledge distillation. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, USA: Curran Associates Inc., 2017. 742-751
[55]	Lin M, Chen Q, Yan S C. Network in network. arXiv: 1312.4400, 2013.
[56]	Ioffe S, Szegedy C. Batch normalization: accelerating deep network training by reducing internal covariate shift. arXiv: 1502.03167, 2015.
[57]	Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z. Rethinking the inception architecture for computer vision. In: Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016. 2818-2826
[58]	Szegedy C, Ioffe S, Vanhoucke V, Alemi A A. Inception-v4, inception-resnet and the impact of residual connections on learning. In: Proceedings of the 31st AAAI Conference on Artificial Intelligence. San Francisco, USA: AAAI, 2017. 12
[59]	Chollet F. Xception: deep learning with depthwise separable convolutions. In: Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI, USA: IEEE, 2017.
[60]	Chang J R, Chen Y S. Batch-normalized maxout network in network. arXiv: 1511.02583, 2015.
[61]	Pang Y W, Sun M L, Jiang X H, Li X L. Convolution in convolution for network in network. IEEE transactions on neural networks and learning systems, 2018, 29(5): 1587-1597 doi: 10.1109/TNNLS.2017.2676130
[62]	Han X M, Dai Q. Batch-normalized mlpconv-wise supervised pre-training network in network. Applied Intelligence, 2018, 48(1): 142-155 doi: 10.1007/s10489-017-0968-2
[63]	Srivastava R K, Greff K, Schmidhuber J. Highway networks. arXiv: 1505.00387, 2015.
[64]	Hochreiter S, Schmidhuber J. Long short-term memory. Neural Computation, 1997, 9(8): 1735-1780 doi: 10.1162/neco.1997.9.8.1735
[65]	Larsson G, Maire M, Shakhnarovich G. Fractalnet: ultra-deep neural networks without residuals. arXiv: 1605.07648, 2016.
[66]	Huang G, Sun Y, Liu Z, Sedra D, Weinberger K Q. Deep networks with stochastic depth. In: Proceedings of the 14th European Conference on Computer Vision. Amsterdam, The Netherlands: Springer, 2016. 646-661
[67]	He K M, Zhang X Y, Ren S Q, Sun J. Identity mappings in deep residual networks. In: Proceedings of the 14th European Conference on Computer Vision. Amsterdam, The Netherlands: Springer, 2016. 630-645
[68]	Xie S N, Girshick R, Dollár P, Tu Z W, He K M. Aggregated residual transformations for deep neural networks. In: Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI, USA: IEEE, 2017. 5987-5995
[69]	LeCun Y, Bengio Y, Hinton G. Deep learning. Nature, 2015, 521(7553): 436-444 doi: 10.1038/nature14539
[70]	Zagoruyko S, Komodakis N. Wide residual networks. arXiv: 1605.07146, 2016.
[71]	Targ S, Almeida D, Lyman K. Resnet in resnet: generalizing residual architectures. arXiv: 1603.08029, 2016.
[72]	Zhang K, Sun M, Han T X, Yuan X F, Guo L R, Liu T. Residual networks of residual networks: multilevel residual networks. IEEE Transactions on Circuits and Systems for Video Technology, 2018, 28(6): 1303-1314 doi: 10.1109/TCSVT.2017.2654543
[73]	Abdi M, Nahavandi S. Multi-residual networks: improving the speed and accuracy of residual networks. arXiv: 1609.05672, 2016.
[74]	Huang G, Liu Z, van der Maaten L, Weinberger K Q. Densely connected convolutional networks. In: Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI, USA: IEEE, 2017.
[75]	张婷, 李玉鑑, 胡海鹤, 张亚红.基于跨连卷积神经网络的性别分类模型.自动化学报, 2016, 42(6): 858-865 doi: 10.16383/j.aas.2016.c150658 Zhang Ting, Li Yu-Jian, Hu Hai-He, Zhang Ya-Hong. A gender classification model based on cross-connected convolutional neural networks. Acta Automatica Sinica, 2016, 42(6): 858-865 doi: 10.16383/j.aas.2016.c150658
[76]	李勇, 林小竹, 蒋梦莹.基于跨连接LeNet-5网络的面部表情识别.自动化学报, 2018, 44(1): 176-182 doi: 10.16383/j.aas.2018.c160835 Li Yong, Lin Xiao-Zhu, Jiang Meng-Ying. Facial expression recognition with cross-connect LeNet-5 network. Acta Automatica Sinica, 2018, 44(1): 176-182 doi: 10.16383/j.aas.2018.c160835
[77]	Howard A G, Zhu M L, Chen B, Kalenichenko D, Wang W J, Weyand T, et al. Mobilenets: efficient convolutional neural networks for mobile vision applications. arXiv: 1704.04861, 2017.
[78]	Sandler M, Howard A, Zhu M L, Zhmoginov A, Chen L C. MobileNetV2: inverted residuals and linear bottlenecks. In: Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, USA: IEEE, 2018. 4510-4520
[79]	Zhang X Y, Zhou X Y, Lin M X, Sun J. ShuffleNet: an extremely efficient convolutional neural network for mobile devices. In: Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, USA: IEEE, 2018.

施引文献

资源附件(0)

访问统计

图(9) / 表(2)

计量

文章访问数: 8448
HTML全文浏览量: 3390
PDF下载量: 2410
被引次数: 0

姓名
邮箱
手机号码
标题
留言内容
验证码

留言板

卷积神经网络结构优化综述

doi: 10.16383/j.aas.c180275

通讯作者:
尹宏鹏重庆大学自动化学院教授.2009年获得重庆大学博士学位.主要研究方向为模式识别与智能系统.本文通信作者.E-mail:yinhongpeng@gmail.com

计量

Structure Optimization of Convolutional Neural Networks: A Survey

Corresponding author: YIN Hong-Peng Professor at the College of Automation, Chongqing University. He received his Ph. D. degree from Chongqing University in 2009. His research interest covers pattern recognition, image processing, and computer vision. Corresponding author of this paper

计量

目录

留言板

卷积神经网络结构优化综述

doi: 10.16383/j.aas.c180275

通讯作者: 尹宏鹏 重庆大学自动化学院教授.2009年获得重庆大学博士学位.主要研究方向为模式识别与智能系统.本文通信作者.E-mail:yinhongpeng@gmail.com

计量

出版历程

Structure Optimization of Convolutional Neural Networks: A Survey

Corresponding author: YIN Hong-Peng Professor at the College of Automation, Chongqing University. He received his Ph. D. degree from Chongqing University in 2009. His research interest covers pattern recognition, image processing, and computer vision. Corresponding author of this paper

计量

出版历程

目录

通讯作者:
尹宏鹏重庆大学自动化学院教授.2009年获得重庆大学博士学位.主要研究方向为模式识别与智能系统.本文通信作者.E-mail:yinhongpeng@gmail.com