轻量化高精度双通道注意力机制模块

doi:10.3778/j.issn.1673-9418.2112052

摘要/Abstract

摘要： 目前大多数注意力机制模块在提高深度学习模型应用精度的同时也带来了模型复杂度的增加。针对这一问题，提出了一种轻量化高精度双通道注意力机制模块（EDCA）。EDCA将特征图分别沿通道、宽度和高度三个方向压缩并重新排列组合，采用一维卷积获取组合后的权重信息，接着将权重信息分割并应用于对应维度以获得特征关注度。在图像分类数据集miniImageNet与目标检测数据集Pascal VOC2007上对EDCA进行充分实验，实验结果表明，与SENet、CBAM、SGE、ECA-Net、Coordinate Attention相比，EDCA所需的计算量以及参数量更少。在miniImageNet数据集上使用ResNet50+EDCA时，Top-1精度较以上方法分别提升0.024 3、0.021 8、0.022 1、0.022 5、0.014 1；在Pascal VOC2007数据集上使用MobileNetV3+YOLOv4+EDCA时，AP50较SENet、CBAM、ECA-Net、Coordinate Attention分别提升0.009 4、0.004 6、0.005 9、0.001 4。

关键词: 注意力机制, 一维卷积, 图像分类, 目标检测

Abstract: At present, most attention mechanism modules improve the application accuracy of deep learning models, but also bring the defect of increased model complexity. In response to this problem, this paper proposes a lightweight and efficient dual-channel attention mechanism module (EDCA). EDCA compresses and rearranges the feature maps in three directions of channel, width, and height. One-dimensional convolution is used to obtain the combined weight information, and then the weight information is segmented and applied to corresponding dimensions to obtain feature attention. This paper conducts a full experiment on EDCA on image classification dataset miniImageNet and target detection dataset Pascal VOC2007. Compared with SENet, CBAM, SGE, ECA-Net and Coordinate Attention, EDCA requires less computation and parameters. When ResNet50+EDCA is used on miniImageNet dataset, the Top-1 accuracy is improved by 0.0243, 0.0218, 0.0221, 0.0225 and 0.0141, respectively. When MobileNetV3+YOLOv4+EDCA is used on Pascal VOC2007 dataset, AP50 is improved by 0.0094, 0.0046, 0.0059 and 0.0014 compared with SENet, CBAM, ECA-Net and Coordinate Attention, respectively.

Key words: attention mechanism, one-dimensional convolution, image classification, object detection

陈晓雷, 卢禹冰, 曹宝宁, 林冬梅. 轻量化高精度双通道注意力机制模块[J]. 计算机科学与探索, 2023, 17(4): 857-867.

CHEN Xiaolei, LU Yubing, CAO Baoning, LIN Dongmei. Lightweight and High-Precision Dual-Channel Attention Mechanism Module[J]. Journal of Frontiers of Computer Science and Technology, 2023, 17(4): 857-867.

参考文献

[1] CHAUDHARI S, MITHAL V, POLATKAN G, et al. An at-tentive survey of attention models[J]. ACM Transactions on Intelligent Systems and Technology, 2021, 12(5): 1-32.
[2] FUKUI H, HIRAKAWA T, YAMASHITA T, et al. Attention branch network: learning of attention mechanism for visual explanation[C]//Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach,Jun 16-20, 2019. Piscataway: IEEE, 2019: 10705-10714.
[3] 刘颖, 雷研博, 范九伦, 等. 基于小样本学习的图像分类技术综述[J]. 自动化学报, 2021, 47(2): 297-315.
LIU Y, LEI Y B, FAN J L, et al. Survey on image classifi-cation technology based on small sample learning[J]. Acta Automatica Sinica, 2021, 47(2): 297-315.
[4] 吴帅, 徐勇, 赵东宁. 基于深度卷积网络的目标检测综述[J]. 模式识别与人工智能, 2018, 31(4): 335-346.
WU S, XU Y, ZHAO D N. Survey of object detection based on deep convolutional network[J]. Pattern Recognition and Artificial Intelligence, 2018, 31(4): 335-346.
[5] 景庄伟, 管海燕, 彭代峰, 等. 基于深度神经网络的图像语义分割研究综述[J]. 计算机工程, 2020, 46(10): 1-17.
JING Z W, GUAN H Y, PENG D F, et al. Survey of research in image semantic segmentation based on deep neural net-work[J]. Computer Engineering, 2020, 46(10): 1-17.
[6] ZHANG Y, LI K, LI K, et al. Image super-resolution using very deep residual channel attention networks[C]//LNCS 11211: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Cham: Springer, 2018: 294-310.
[7] CHU Q, OUYANG W, LI H, et al. Online multi-object trac-king using CNN-based single object tracker with spatial-temporal attention mechanism[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 4846-4855.
[8] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]//Proceedings of the 2018 IEEE Conference on Computer Vi-sion and Pattern Recognition, Salt Lake City, Jun 18-23, 2018. Washington: IEEE Computer Society, 2018: 7132-7141.
[9] HU J, SHEN L, ALBANIE S, et al. Gather-excite: exploiting feature context in convolutional neural networks[J]. arXiv:1810.12348, 2018.
[10] PARK J, WOO S, LEE J Y, et al. BAM: bottleneck attention module[J]. arXiv:1807.06514, 2018.
[11] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]//LNCS 11211: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Cham: Springer, 2018: 3-19.
[12] FU J, LIU J, TIAN H, et al. Dual attention network for scene segmentation[C]//Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, Jun 16-20, 2019. Piscataway: IEEE, 2019: 3146-3154.
[13] LI X, HU X, YANG J. Spatial group-wise enhance: improving semantic feature learning in convolutional networks[J]. arXiv:1905.09646, 2019.
[14] ZHANG Q L, YANG Y B. SA-NET: shuffle attention for deep convolutional neural networks[C]//Proceedings of the 2021 IEEE International Conference on Acoustics, Speech and Signal Processing, Toronto, Jun 6-11, 2021. Piscataway: IEEE, 2021: 2235-2239.
[15] WANG Q L, WU B G, ZHU P F, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]//Proceedings of the 2020 IEEE/CVF Conference on Com-puter Vision and Pattern Recognition, Seattle, Jun 13-19, 2020. Piscataway: IEEE, 2020: 11531-11539.
[16] HOU Q, ZHOU D, FENG J. Coordinate attention for efficient mobile network design[C]//Proceedings of the 2021 IEEE Conference on Computer Vision and Pattern Recognition, Jun 19-25, 2021. Piscataway: IEEE, 2021: 13713-13722.
[17] MA X, GUO J, TANG S, et al. DCANet: learning connected attentions for convolutional neural networks[J]. arXiv:2007.05099, 2020.
[18] LI X, WANG W H, HU X L, et al. Selective kernel networks[C]//Proceedings of the 2019 IEEE Conference on Com-puter Vision and Pattern Recognition, Long Beach, Jun 16-20, 2019. Piscataway: IEEE, 2019: 510-519.
[19] SAGAR A. DMSANet: dual multi scale attention network[J]. arXiv:2106.08382, 2021.
[20] CHEN Y, KALANTIDIS Y, LI J, et al. A2-Nets: double at-tention networks[J]. arXiv:1810.11579, 2018.
[21] WANG F, JIANG M Q, QIAN C, et al. Residual attention network for image classification[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer So-ciety, 2017: 6450-6458.
[22] ZHU X Z, CHENG D Z, ZHANG Z, et al. An empirical study of spatial attention mechanisms in deep networks[C]//Pro-ceedings of the 2019 IEEE/CVF International Conference on Computer Vision, Seoul, Oct 27-Nov 2, 2019. Piscataway: IEEE, 2019: 6688-6697.
[23] CAI Q, PAN Y W, YAO T, et al. Memory matching networks for one-shot image recognition[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recog-nition, Salt Lake City, Jun 18-22, 2018. Piscataway: IEEE, 2018: 4080-4088.
[24] EVERINGHAM M, ESLAMI S M A, VAN GOOL L, et al. The pascal visual object classes challenge: a retrospective[J]. International Journal of Computer Vision, 2015, 111(1): 98-136.
[25] HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recogni-tion, Las Vegas, Jun 27-30, 2016. Washington: IEEE Com-puter Society, 2016: 770-778.
[26] HAN K, WANG Y H, TIAN Q, et al. GhostNet: more features from cheap operations[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recogni-tion, Seattle, Jun 13-19, 2020. Piscataway: IEEE, 2020: 1577-1586.
[27] BOCHKOVSKIY A, WANG C Y, LIAO H Y M. YOLOv4: optimal speed and accuracy of object detection[J]. arXiv:2004.10934, 2020.
[28] HOWARD A, SANDLER M, CHU G, et al. Searching for MobileNetV3[C]//Proceedings of the 2019 IEEE/CVF Inter-national Conference on Computer Vision, Seoul, Oct 27-Nov 2, 2019. Piscataway: IEEE, 2019: 1314-1324.