[1]陈琳琳,朱惠娟,朱 俊,等.基于卷积神经网络的多尺度注意力图像分类模型[J].南京理工大学学报(自然科学版),2020,44(06):669-675.[doi:10.14177/j.cnki.32-1397n.2020.44.06.005]
 Chen Linlin,Zhu Huijuan,Zhu Jun,et al.Multiscale attention model for image classificationbased on convolutional neural network[J].Journal of Nanjing University of Science and Technology,2020,44(06):669-675.[doi:10.14177/j.cnki.32-1397n.2020.44.06.005]
点击复制

基于卷积神经网络的多尺度注意力图像分类模型()
分享到:

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

卷:
44卷
期数:
2020年06期
页码:
669-675
栏目:
出版日期:
2020-12-31

文章信息/Info

Title:
Multiscale attention model for image classificationbased on convolutional neural network
文章编号:
1005-9830(2020)06-0669-07
作者:
陈琳琳12朱惠娟1朱 俊1王晓瞳1
1.南京理工大学紫金学院 计算机学院,江苏 南京 210023; 2.南京理工大学 计算机科学与工程学院,江苏 南京 210094
Author(s):
Chen Linlin12Zhu Huijuan1Zhu Jun1Wang Xiaotong1
1.Computer College,Nanjing University of Science and Technology Zijin College,Nanjing 210023,China; 2.School of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094,China)
关键词:
卷积神经网络 多尺度 注意力 图像 分类 残差网络
Keywords:
convolutional neural network multiscale attention images classification residual network
分类号:
TP391
DOI:
10.14177/j.cnki.32-1397n.2020.44.06.005
摘要:
为解决传统的图像分类方法精度低的问题,该文提出了一种基于卷积神经网络的图像分类网络模型。首先,提出了一种基于注意力机制的特征注意力模块(FAM),提取不同权重的图像特征。然后,提出了基于FAM的多尺度注意力卷积神经网络(MSACNN),通过3个FAM块提取不同尺度下的精确的图像特征进行分类。将MSACNN与3种典型的卷积神经网络LeNet-5、AlexNet以及残差网络(ResNet)在MNIST数据集上进行了对比,结果表明,MSACNN的分类精度和稳定性效果最好。
Abstract:
In order to solve the problem of low accuracy of traditional image classification methods,an image classification network based on convolutional neural network is proposed here. Firstly,a kind of feature attention module(FAM)based on attention mechanism is proposed to extract image feature maps with different weights. Next,three FAM blocks are used in the network to extract accurate image feature maps at different scales for the following image classification,and a multiscale attention convolutional neural network(MSACNN)is proposed. Finally,the proposed network is compared with 3 typical convolutional neural networks of LeNet-5,AlexNet and residual network(ResNet)on Mixed National Institute of Standards and Technology(MNIST)dataset,and the experimental results show that the MSACNN achieves the best results in both classification accuracy and stability.

参考文献/References:

[1] 王彦翔,张艳,杨成娅,等. 基于深度学习的农作物病害图像识别技术进展[J]. 浙江农业学报,2019,31(4):669-676.
Wang Yanxiang,Zhang Yan,Yang Chengya,et al. Advances in new nondestructive detection and identification techniques of crop diseases based on deep learning[J]. Acta Agriculturae Zhejiangensis,2019,31(4):669-676.
[2]孙雨萌. 深度学习在计算机视觉分析中的应用分析[J]. 中国新通信,2018,20(23):169-171.
Sun Yumeng. Application analysis of deep learning in computer vision analysis[J]. China New Telecommunications,2018,20(23):169-171.
[3]孙康,李千目,李德强. 基于级联卷积神经网络的人脸检测算法[J]. 南京理工大学学报,2018,42(1):40-47.
Sun Kang,Li Qianmu,Li Deqiang. Face detection algorithm based on cascaded convolutional neural network[J]. Journal of Nanjing University of Science and Technology,2018,42(1):40-47.
[4]张新钰,高洪波,赵建辉,等. 基于深度学习的自动驾驶技术综述[J]. 清华大学学报(自然科学版),2018,58(4):438-444.
Zhang Xinyu,Gao Hongbo,Zhao Jianhui,et al. Overview of deep learning intelligent driving methods[J]. Journal of Tsinghua University(Science and Technology),2018,58(4):438-444.
[5]梁蒙蒙,周涛,张飞飞,等. 卷积神经网络及其在医学图像分析中的应用研究[J]. 生物医学工程学杂志,2018,35(6):977-985.
Liang Mengmeng,Zhou Tao,Zhang Feifei,et al. Research on convolutional neural network and its application on medical image[J]. Journal of Biomedical Engineering,2018,35(6):977-985.
[6]Lecun Y,Bottou L,Bengia Y,et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE,1998,86(11):2278-2324.
[7]Krizhevsky A,Sutskever I,Hinton G E. ImageNet classification with deep convolutional neural networks[J]. Advances in Neural Information Processing Systems,2012,25(2):4824-4832.
[8]Zeiler M D,Fergus R. Visualizing and understanding convolutional networks[EB/OL]. https://arxiv. org/pdf/1311. 2901,2020-12-03.
[9]Simonyan K,Zisserman A. Very deep convolutional networks for large-scale image recognition[EB/OL]. https://arxiv. org/pdf/1409. 1556. pdf,2020-12-03.
[10]Szegedy C,Liu Wei,Jia Yangqing,et al. Going deeper with convolutions[C]//Proceedings of the 2014 IEEE Conference on Computer Vision and Pattern Recognition. Washington DC,USA:IEEE,2014:1-9.
[11]He Kaiming,Zhang Xiangyu,Ren Shaoqing,et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Washington DC,USA:IEEE,2016:770-778.
[12]Hu Jie,Shen Li,Albanie S,et al. Squeeze-and-excitation networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2020,42(8):2011-2023.
[13]Bahdanau D,Cho K H,Bengio Y. Neural machine translation by jointly learning to align and translate[EB/OL]. https://arxiv. org/pdf/1409. 0473. pdf,2020-12-03.
[14]Chen Long,Zhang Hanwang,Xiao Jun,et al. SCA-CNN:Spatial and channel-wise attention in convolutional networks for image captioning[EB/OL]. https://arxiv. org/pdf/1611. 05594. pdf,2020-12-03.
[15]Wang Fei,Jiang Mengqing,Qian Chen. Residual attention network for image classification[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu,HI,USA:IEEE,2017:3156-3164.
[16]Woo S,Park J,Lee J Y,et al. CBAM:Convolutional block attention module[EB/OL]https://arxiv. org/pdf/1807. 06521. pdf,2020-12-03.
[17]Fu Jun,Liu Jing,Tian Haijie,et al. Dual attention network for scene segmentation[C]//Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach,CA,USA:IEEE,2019:3141-3149.
[18]何柔萤,徐建. 基于注意力卷积神经网络的工作票专家推荐方法[J]. 南京理工大学学报,2019,43(1):13-21,47.
He Rouying,Xu Jian. Expert recommendation for trouble tickets using attention-based CNN model[J]. Journal of Nanjing University of Science and Technology,2019,43(1):13-21,47.
[19]姜枫,张丽红. 基于随机Dropout卷积神经网络的人体行为识别方法研究[J]. 测试技术学报,2016,30(1):17-22.
Jiang Feng,Zhang Lihong. Research of human action recognition method based on random Dropout convolutional neural network[J]. Journal of Test and Measurement Technology,2016,30(1):17-22.

相似文献/References:

[1]葛 婷,詹天明,牟善祥.基于多核协同表示分类的脑肿瘤分割算法[J].南京理工大学学报(自然科学版),2019,43(05):578.[doi:10.14177/j.cnki.32-1397n.2019.43.05.006]
 Tianming,Mu Shanxiang.Brain tumor segmentation algorithm based on multi-kernelcollaborative representation classificationGe Ting1,2,Zhan[J].Journal of Nanjing University of Science and Technology,2019,43(06):578.[doi:10.14177/j.cnki.32-1397n.2019.43.05.006]
[2]郭业才,朱文军.基于深度卷积神经网络的运动模糊去除算法[J].南京理工大学学报(自然科学版),2020,44(03):303.[doi:10.14177/j.cnki.32-1397n.2020.44.03.007]
 Guo Yecai,Zhu Wenjun.Motion deblurring algorithm based on deep convolutional neural network[J].Journal of Nanjing University of Science and Technology,2020,44(06):303.[doi:10.14177/j.cnki.32-1397n.2020.44.03.007]

备注/Memo

备注/Memo:
收稿日期:2020-09-24 修回日期:2020-10-18
基金项目:江苏省高等学校自然科学研究项目(18KJB520023; 20KJB520035); 江苏高校“青蓝工程”培养项目; 南京理工大学紫金学院科研项目(2020ZRKX0401001)
作者简介:陈琳琳(1981-),女,博士生,讲师,主要研究方向:深度学习、高光谱图像处理,E-mail:chenlinlin606@njust.edu.cn。
引文格式:陈琳琳,朱惠娟,朱俊,等. 基于卷积神经网络的多尺度注意力图像分类模型[J]. 南京理工大学学报,2020,44(6):669-675.
投稿网址:http://zrxuebao.njust.edu.cn
更新日期/Last Update: 2020-12-30