基于跨通道交叉融合和跨模块连接的轻量级卷积神经网络

doi:10.11772/j.issn.1001-9081.2020060882

计算机应用 ›› 2020, Vol. 40 ›› Issue (12): 3451-3457.DOI: 10.11772/j.issn.1001-9081.2020060882

• 2020年中国粒计算与知识发现学术会议(CGCKD 2020) • 上一篇下一篇

基于跨通道交叉融合和跨模块连接的轻量级卷积神经网络

陈力, 丁世飞, 于文家

中国矿业大学计算机科学与技术学院, 江苏徐州 221116

收稿日期:2020-06-19 修回日期:2020-08-24 出版日期:2020-12-10 发布日期:2020-10-20
通讯作者: 丁世飞(1963-),男,山东青岛人,教授,博士,CCF会员,主要研究方向:人工智能、机器学习、深度强化学习。dingsf@cumt.edu.cn
作者简介:陈力(1993-),男,山东邹城人,硕士研究生,主要研究方向:深度学习、图像处理;于文家(1994-),男,辽宁本溪人,硕士研究生,主要研究方向:深度学习
基金资助:
国家自然科学基金资助项目（61672522，61976216，61379101）。

Lightweight convolutional neural network based on cross-channel fusion and cross-module connection

CHEN Li, DING Shifei, YU Wenjia

School of Computer Science and Technology, China University of Mining and Technology, Xuzhou Jiangsu 221116, China

Received:2020-06-19 Revised:2020-08-24 Online:2020-12-10 Published:2020-10-20
Supported by:
This work is partially supported by the National Natural Science Foundation of China （61672522， 61976216， 61379101）.

摘要/Abstract

摘要： 针对传统卷积神经网络参数量过多、计算复杂度高的问题，提出了基于跨通道交叉融合和跨模块连接的轻量级卷积神经网络架构C-Net。首先，提出了跨通道交叉融合的方法，它在一定程度上克服了分组卷积中各分组之间存在缺乏信息流动的问题，简单高效地实现了不同分组之间的信息通信；其次，提出了一种跨模块连接的方法，它克服了传统轻量级架构中各基本构建块之间彼此独立的缺点，实现了同一阶段内具有相同分辨率特征映射的不同模块之间的信息融合，从而增强了特征提取能力；最后，基于提出的两种方法设计了一种新型的轻量级卷积神经网络架构C-Net。C-Net在Food_101数据集上的准确率为69.41%，在Caltech_256数据集上的准确率为63.93%。实验结果表明，与目前先进的轻量级卷积神经网络模型相比，C-Net降低了存储开销和计算复杂度。在Cifar_10数据集上的消融实验验证了所提出的两种方法的有效性。

关键词: 卷积神经网络, 轻量级, 分组卷积, 跨通道交叉融合, 快捷连接, 跨模块连接

Abstract: In order to solve the problems of too many parameters and high computational complexity of traditional convolutional neural networks, a lightweight convolutional neural network architecture named C-Net based on cross-channel fusion and cross-module connection was proposed. Firstly, a method called cross-channel fusion was proposed. With it, the shortcoming of lacking information flow between different groups of grouped convolution was solved to a certain extent, and the information communication between different groups was realized efficiently and easily. Then, a method called cross-module connection was proposed. With it, the shortcoming that the basic building blocks in the traditional lightweight architecture were independent to each other was overcome, and the information fusion between different modules with the same resolution feature mapping within the same stage was achieved, enhancing the feature extraction capability. Finally, a novel lightweight convolutional neural network architecture C-Net was designed based on the two proposed methods. The accuracy of C-Net on the Food_101 dataset is 69.41%, and the accuracy of C-Net on the Caltech_256 dataset is 63.93%. Experimental results show that C-Net reduces the memory cost and computational complexity in comparison with the state-of-the-art lightweight convolutional neural network models. The ablation experiment verifies the effectiveness of the two proposed methods on the Cifar_10 dataset.

Key words: Convolutional Neural Network (CNN), lightweight, grouped convolution, cross-channel fusion, shortcut connection, cross-module connection

中图分类号:

TP391.4

陈力, 丁世飞, 于文家. 基于跨通道交叉融合和跨模块连接的轻量级卷积神经网络[J]. 计算机应用, 2020, 40(12): 3451-3457.

CHEN Li, DING Shifei, YU Wenjia. Lightweight convolutional neural network based on cross-channel fusion and cross-module connection[J]. Journal of Computer Applications, 2020, 40(12): 3451-3457.

参考文献

[1] HINTON G E, SALAKHUTDINOV R R. Reducing the dimensionality of data with neuralnetworks[J]. Science,2006, 313(5786):504-507.
[2] LECUN Y, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE,1998,86(11):2278-2324.
[3] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neuralnetworks[C]//Proceedings of the 25th International Conference on Neural Information Processing Systems. Red Hook:Curran Associates Inc.,2012:1097-1105.
[4] RUSSAKOVSKY O,DENG J,SU H,et al. ImageNet large scale visual recognition challenge[J]. International Journal of Computer Vision,2015,115(3):211-252.
[5] SIMONYAN K,ZISSERMAN A. Very deep convolutionalnetworks for large-scale image recognition[EB/OL].[2020-03-04]. https://arxiv.org/pdf/1409.1556.pdf.
[6] DUMOULIN V,VISIN F. A guide to convolution arithmetic for deep learning[EB/OL].[2020-03-23]. https://arxiv.org/pdf/1603.07285.pdf.
[7] SZEGEDY C, LIU W, JIA Y, et al. Going deeper with convolutions[C]//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2015:1-9.
[8] HE K,ZHANG X,REN S,et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2016:770-778.
[9] HUANG G, LIU Z, VAN DER MAATEN L, et al. Densely connected convolutionalnetworks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2017:2261-2269.
[10] HU J,SHEN L,SUN G. Squeeze-and-excitationnetworks[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2018:7132-7141.
[11] 纪荣嵘, 林绍辉, 晁飞, 等. 深度神经网络压缩与加速综述[J]. 计算机研究与发展, 2018, 55(9):1871-1888.(JI R R,LIN S H, CHAO F, et al. Deep neuralnetworkcompression and acceleration:a review[J]. Journal of Computer Research and Development,2018,55(9):1871-1888.)
[12] SETIONO R,LIU H. Neural-network feature selector[J]. IEEE Transactions on Neural Networks,1997,8(3):654-662.
[13] ZHANG Y,JIANG Z,DAVIS L S. Learning structured low-rank representations for image classification[C]//Proceedings of the 2013 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2013:676-683.
[14] PENG C,ZHANG X,YU G,et al. Large kernel matters-improve semantic segmentation by global convolutionalnetwork[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2017:1743-1751.
[15] BUCILUǍ C, CARUANA R, NICULESCU-MIZIL A. Modelcompression[C]//Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York:ACM,2006:535-541.
[16] GAO L,CHEN P Y,YU S. Demonstration of convolution kernel operation on resistive cross-point array[J]. IEEE Electron Device Letters,2016,37(7):870-873.
[17] CHOLLET F. Xception:deep learning with depthwise separable convolutions[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2017:1800-1807.
[18] LIN M,CHEN Q,YAN S. Network innetwork[EB/OL].[2020-03-16]. https://arxiv.org/pdf/1312.4400.pdf.
[19] RUMELHART D E,HINTON G E,WILLIAMS R J,et al. Learning representations by back-propagating errors[J]. Nature, 1986,323(6088):533-536.
[20] HOCHREITER S. The vanishing gradient problem during learning recurrent neuralnets and problem solutions[J]. International Journal of Uncertainty, Fuzziness and Knowledge-Based, Systems,1998,6(2):107-116.
[21] GLOROT X,BENGIO Y. Understanding the difficulty of training deep feedforward neuralnetworks[J]. Journal of Machine Learning Research,2010,9:249-256.
[22] ZHANG X,ZHOU X,LIN M,et al. ShuffleNet:an extremely efficient convolutional neuralnetwork for mobile devices[EB/OL].[2020-03-04]. https://arxiv.org/pdf/1707.01083.pdf.
[23] MA N,ZHANG X,ZHENG H,et al. ShuffleNet V2:practical guidelines for efficient CNN architecture design[C]//Proceedings of the 2018 European Conference on Computer Vision,LNCS 11218. Cham:Springer,2018:122-138.
[24] XIE S,GIRSHICK R,DOLLÁR P,et al. Aggregated residual transformations for deep neuralnetworks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2017:5987-5995.
[25] IANDOLA F N,HAN S,MOSKEWICZ M W,et al. SqueezeNet:AlexNet-level accuracy with 50x fewer parameters and <0.5 MB model size[EB/OL].[2020-03-24]. https://arxiv.org/pdf/1602.07360.pdf.
[26] HOWARD A G,ZHU M,CHEN B,et al. MobileNets:efficient convolutional neuralnetworks for mobile vision applications[EB/OL].[2020-03-17]. https://arxiv.org/pdf/1704.04861.pdf.
[27] SANDLER M,HOWARD A,ZHU M,et al. MobileNetV2:inverted residuals and linear bottlenecks[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2018:4510-4520.
[28] LEE C Y,GALLAGHER P W,TU Z. Generalizing pooling functions in convolutional neuralnetworks:mixed,gated,and tree[C]//Proceedings of the 19th International Conference on Artificial Intelligence and Statistics. Cambridge:MIT Press, 2016:464-472.
[29] SCHERER D,MÜLLER A,BEHNKE S. Evaluation of pooling operations in convolutional architectures for object recognition[C]//Proceedings of the 20th International Conference on Artificial Neural Networks,LNCS 6354. Berlin:Springer,2010:92-101.
[30] HAN D,KIM J,KIM J. Deep pyramidal residualnetworks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2017:6307-6315.
[31] ZHANG S,ZHANG S,ZHANG C,et al. Cucumber leaf disease identification with global pooling dilated convolutional neuralnetwork[J]. Computers and Electronics in Agriculture,2019, 162:422-430.
[32] SRIVASTAVA N, HINTON G, KRIZHEVSKY A, et al. Dropout:a simple way to prevent neuralnetworks from overfitting[J]. Journal of Machine Learning Research,2014,15(1):1929-1958.
[33] IOFFE S,SZEGEDY C. Batch normalization:accelerating deepnetwork training by reducing internal covariate shift[C]//Proceedings of the 32nd International Conference on Machine Learning. New York:ACM,2015:448-456.
[34] KETKAR N. Introduction to PyTorch[M]//Deep Learning with Python:A Hands-on Introduction. Berkeley,CA:Apress,2017:195-208.
[35] 张蕊, 李锦涛. 基于深度学习的场景分割算法研究综述[J]. 计算机研究与发展, 2020, 57(4):859-875.(ZHANG R,LI J T. A survey on algorithm research of scene parsing based on deep learning[J]. Journal of Computer Research and Development, 2020,57(4):859-875.)
[36] 黄继鹏, 史颖欢, 高阳. 面向小目标的多尺度Faster-RCNN检测算法[J]. 计算机研究与发展, 2019, 56(2):319-327.(HUANG J P,SHI Y H,GAO Y. Multi-scale faster-RCNN algorithm for small object detection[J]. Journal of Computer Research and Development,2019,56(2):319-327.)

基于跨通道交叉融合和跨模块连接的轻量级卷积神经网络

Lightweight convolutional neural network based on cross-channel fusion and cross-module connection

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	宋中山, 梁家锐, 郑禄, 刘振宇, 帖军. 基于双向门控尺度特征融合的遥感场景分类[J]. 计算机应用, 2021, 41(9): 2726-2735.
[2]	李康康, 张静. 基于注意力机制的多层次编码和解码的图像描述模型[J]. 计算机应用, 2021, 41(9): 2504-2509.
[3]	张永斌, 常文欣, 孙连山, 张航. 基于字典的域名生成算法生成域名的检测方法[J]. 计算机应用, 2021, 41(9): 2609-2614.
[4]	赵宏, 孔东一. 图像特征注意力与自适应注意力融合的图像内容中文描述[J]. 计算机应用, 2021, 41(9): 2496-2503.
[5]	徐江浪, 李林燕, 万新军, 胡伏原. 结合目标检测的室内场景识别方法[J]. 计算机应用, 2021, 41(9): 2720-2725.
[6]	牟长宁, 王海鹏, 周丕宇, 侯鑫行. 基于图卷积神经网络的串联质谱从头测序[J]. 计算机应用, 2021, 41(9): 2773-2779.
[7]	王贺兵, 张春梅. 基于非对称卷积-压缩激发-次代残差网络的人脸关键点检测[J]. 计算机应用, 2021, 41(9): 2741-2747.
[8]	曹玉红, 徐海, 刘荪傲, 王紫霄, 李宏亮. 基于深度学习的医学影像分割研究综述[J]. 计算机应用, 2021, 41(8): 2273-2287.
[9]	秦斌斌, 彭良康, 卢向明, 钱江波. 司机分心驾驶检测研究进展[J]. 计算机应用, 2021, 41(8): 2330-2337.
[10]	黄程程, 董霄霄, 李钊. 基于二维Winograd算法的深流水线5×5卷积方法[J]. 计算机应用, 2021, 41(8): 2258-2264.
[11]	曾祥银, 郑伯川, 刘丹. 基于深度卷积神经网络和聚类的左右轨道线检测[J]. 计算机应用, 2021, 41(8): 2324-2329.
[12]	吴则举, 焦翠娟, 陈亮. 基于改进Faster R-CNN的轮胎缺陷检测方法[J]. 计算机应用, 2021, 41(7): 1939-1946.
[13]	杨粟, 欧阳智, 杜逆索. 基于相关度距离的无监督并行哈希图像检索[J]. 计算机应用, 2021, 41(7): 1902-1907.
[14]	高钦泉, 黄炳城, 刘文哲, 童同. 基于改进CenterNet的竹条表面缺陷检测方法[J]. 计算机应用, 2021, 41(7): 1933-1938.
[15]	冯兴杰, 张天泽. 基于分组卷积进行特征融合的全景分割算法[J]. 计算机应用, 2021, 41(7): 2054-2061.