Journal of Computer Applications ›› 2023, Vol. 43 ›› Issue (3): 685-691.DOI: 10.11772/j.issn.1001-9081.2022010032
Special Issue: 人工智能
• Artificial intelligence • Previous Articles Next Articles
Zhenliang LI1, Bo LI2()
Received:
2022-01-11
Revised:
2022-03-13
Accepted:
2022-03-22
Online:
2022-04-11
Published:
2023-03-10
Contact:
Bo LI
About author:
LI Zhenliang, born in 1997, M. S. candidate. His research interests include deep learning, object detection.通讯作者:
李波
作者简介:
李振亮(1997—),男,河南许昌人,硕士研究生,主要研究方向:深度学习、目标检测CLC Number:
Zhenliang LI, Bo LI. Improved method of convolution neural network based on matrix decomposition[J]. Journal of Computer Applications, 2023, 43(3): 685-691.
李振亮, 李波. 基于矩阵分解的卷积神经网络改进方法[J]. 《计算机应用》唯一官方网站, 2023, 43(3): 685-691.
Add to citation manager EndNote|Ris|BibTeX
URL: https://www.joca.cn/EN/10.11772/j.issn.1001-9081.2022010032
模型 | 准确率/% | 训练用时/s | 推理用时/ms |
---|---|---|---|
VGG11 | 85.58 | 4 004 | 4 596 |
VGG11+TQRD | 87.42 | 4 173 | 4 597 |
VGG11+RSVD | 86.75 | 4 367 | 4 489 |
VGG13 | 88.22 | 4 929 | 4 770 |
VGG13+TQRD | 89.33 | 5 402 | 4 795 |
VGG13+RSVD | 88.87 | 5 439 | 4 785 |
VGG16 | 86.24 | 5 762 | 5 025 |
VGG16+TQRD | 87.19 | 6 242 | 5 018 |
VGG16+RSVD | 86.40 | 6 579 | 5 098 |
VGG19 | 86.34 | 6 474 | 5 341 |
VGG19+TQRD | 87.39 | 7 094 | 5 364 |
VGG19+RSVD | 87.21 | 7 522 | 5 365 |
Tab. 1 Improvement effect comparison on VGG models
模型 | 准确率/% | 训练用时/s | 推理用时/ms |
---|---|---|---|
VGG11 | 85.58 | 4 004 | 4 596 |
VGG11+TQRD | 87.42 | 4 173 | 4 597 |
VGG11+RSVD | 86.75 | 4 367 | 4 489 |
VGG13 | 88.22 | 4 929 | 4 770 |
VGG13+TQRD | 89.33 | 5 402 | 4 795 |
VGG13+RSVD | 88.87 | 5 439 | 4 785 |
VGG16 | 86.24 | 5 762 | 5 025 |
VGG16+TQRD | 87.19 | 6 242 | 5 018 |
VGG16+RSVD | 86.40 | 6 579 | 5 098 |
VGG19 | 86.34 | 6 474 | 5 341 |
VGG19+TQRD | 87.39 | 7 094 | 5 364 |
VGG19+RSVD | 87.21 | 7 522 | 5 365 |
模型 | 准确率/% | 训练用时/s | 推理用时/ms |
---|---|---|---|
ResNet18 | 87.00 | 9 172 | 5 980 |
ResNet18+TQRD | 87.66 | 9 616 | 6 091 |
ResNet18+RSVD | 87.61 | 10 543 | 6 063 |
ResNet34 | 87.64 | 14 603 | 7 357 |
ResNet34+TQRD | 89.25 | 15 605 | 7 468 |
ResNet34+RSVD | 88.27 | 15 989 | 7 348 |
ResNet50a | 85.96 | 21 727 | 10 764 |
ResNet50a+TQRD | 86.29 | 21 254 | 10 774 |
ResNet50a+RSVD | 86.04 | 21 712 | 10 738 |
ResNet50b | 85.96 | 21 727 | 10 764 |
ResNet50b+TQRD | 86.81 | 21 334 | 11 107 |
ResNet50b+RSVD | 87.11 | 22 054 | 11 013 |
Tab. 2 Improved effects comparison on ResNet models
模型 | 准确率/% | 训练用时/s | 推理用时/ms |
---|---|---|---|
ResNet18 | 87.00 | 9 172 | 5 980 |
ResNet18+TQRD | 87.66 | 9 616 | 6 091 |
ResNet18+RSVD | 87.61 | 10 543 | 6 063 |
ResNet34 | 87.64 | 14 603 | 7 357 |
ResNet34+TQRD | 89.25 | 15 605 | 7 468 |
ResNet34+RSVD | 88.27 | 15 989 | 7 348 |
ResNet50a | 85.96 | 21 727 | 10 764 |
ResNet50a+TQRD | 86.29 | 21 254 | 10 774 |
ResNet50a+RSVD | 86.04 | 21 712 | 10 738 |
ResNet50b | 85.96 | 21 727 | 10 764 |
ResNet50b+TQRD | 86.81 | 21 334 | 11 107 |
ResNet50b+RSVD | 87.11 | 22 054 | 11 013 |
VGG11模块 | TQRD | RSVD | ||||
---|---|---|---|---|---|---|
C1 | C2 | C3 | C4 | C5 | ||
— | — | — | — | — | 85.58 | 85.58 |
√ | — | — | — | — | 86.83 | 86.81 |
— | √ | — | — | — | 86.07 | 85.52 |
— | — | √ | — | — | 86.66 | 85.65 |
— | — | — | √ | 85.65 | 84.66 | |
— | — | — | — | √ | 85.05 | 85.05 |
√ | √ | — | — | — | 86.47 | 86.27 |
√ | √ | √ | — | — | 87.4 | 87.25 |
√ | √ | √ | √ | 87.57 | 86.54 | |
√ | √ | √ | √ | √ | 87.42 | 86.75 |
Tab. 3 Accuracy comparison of different modules in VGG11
VGG11模块 | TQRD | RSVD | ||||
---|---|---|---|---|---|---|
C1 | C2 | C3 | C4 | C5 | ||
— | — | — | — | — | 85.58 | 85.58 |
√ | — | — | — | — | 86.83 | 86.81 |
— | √ | — | — | — | 86.07 | 85.52 |
— | — | √ | — | — | 86.66 | 85.65 |
— | — | — | √ | 85.65 | 84.66 | |
— | — | — | — | √ | 85.05 | 85.05 |
√ | √ | — | — | — | 86.47 | 86.27 |
√ | √ | √ | — | — | 87.4 | 87.25 |
√ | √ | √ | √ | 87.57 | 86.54 | |
√ | √ | √ | √ | √ | 87.42 | 86.75 |
数据集 | 模型 | Baseline | TQRD | RSVD |
---|---|---|---|---|
Fashion-MNIST | VGG11 | 92.75 | 93.29 | 92.88 |
ResNet18 | 93.92 | 94.14 | 94.16 | |
EMNIST Balanced | VGG11 | 89.20 | 89.40 | 89.37 |
ResNet18 | 89.17 | 89.32 | 89.44 | |
CIFAR-100 | VGG11 | 54.43 | 57.75 | 55.95 |
ResNet18 | 58.51 | 60.29 | 59.20 |
Tab. 4 Classification accuracy comparison on different datasets
数据集 | 模型 | Baseline | TQRD | RSVD |
---|---|---|---|---|
Fashion-MNIST | VGG11 | 92.75 | 93.29 | 92.88 |
ResNet18 | 93.92 | 94.14 | 94.16 | |
EMNIST Balanced | VGG11 | 89.20 | 89.40 | 89.37 |
ResNet18 | 89.17 | 89.32 | 89.44 | |
CIFAR-100 | VGG11 | 54.43 | 57.75 | 55.95 |
ResNet18 | 58.51 | 60.29 | 59.20 |
1 | KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[C]// Proceedings of the 25th International Conference on Neural Information Processing Systems. Red Hook, NY: Curran Associates Inc., 2012: 1097-1105. |
2 | SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition [EB/OL]. (2015-04-10) [2021-12-26]. . |
3 | 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 Recognition. Piscataway: IEEE, 2016: 770-778. 10.1109/cvpr.2016.90 |
4 | HOWARD A G, ZHU M L, CHEN B, et al. MobileNets: efficient convolutional neural networks for mobile vision applications [EB/OL]. (2017-04-17) [2021-11-22]. . 10.48550/arXiv.1704.04861 |
5 | REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection [C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 779-788. 10.1109/cvpr.2016.91 |
6 | REN S Q, HE K M, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[C]// Proceedings of the 28th International Conference on Neural Information Processing Systems. Cambridge: MIT Press, 2015: 91-99. |
7 | 张瑶,卢焕章,张路平,等.基于深度学习的视觉多目标跟踪算法综述[J].计算机工程与应用,2021,57(13):55-66. |
ZHANG Y, LU H Z, ZHANG L P, et al. Overview of visual multi-object tracking algorithms with deep learning [J]. Computer Engineering and Applications, 2021, 57(13): 55-66. | |
8 | 徐辉,祝玉华,甄彤,等.深度神经网络图像语义分割方法综述[J].计算机科学与探索,2021,15(1):47-59. 10.3778/j.issn.1673-9418.2004039 |
XU H, ZHU Y H, ZHEN T, et al. Survey of image semantic segmentation methods based on deep neural network [J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(1): 47-59. 10.3778/j.issn.1673-9418.2004039 | |
9 | 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. 10.1007/s11263-015-0816-y |
10 | ALLEN-ZHU Z, LI Y Z, SONG Z. A convergence theory for deep learning via over-parameterization[C]// Proceedings of the 36th International Conference on Machine Learning. New York: JMLR.org, 2019: 242-252. |
11 | ARORA S, COHEN N, HAZAN E. On the optimization of deep networks: implicit acceleration by overparameterization[C]// Proceedings of the 35th International Conference on Machine Learning. New York: JMLR.org, 2018: 244-253. |
12 | COSNARD M, MULLER J M, ROBERT Y. Parallel QR decomposition of a rectangular matrix [J]. Numerische Mathematik, 1986, 48(2): 239-249. 10.1007/bf01389871 |
13 | KLEMA V, LAUB A. The singular value decomposition: its computation and some applications [J]. IEEE Transactions on Automatic Control, 1980, 25(2): 164-176. 10.1109/tac.1980.1102314 |
14 | SRIVASTAVA R K, GREFF K, SCHMIDHUBER J. Highway networks [EB/OL]. (2015-11-03) [2021-10-15]. . |
15 | SZEGEDY C, LIU W, JIA Y Q, et al. Going deeper with convolutions [C]// Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2015: 1-9. 10.1109/cvpr.2015.7298594 |
16 | ZHANG X Y, ZHOU X Y, LIN M X, et al. ShuffleNet: an extremely efficient convolutional neural network for mobile devices [C]// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 6848-6856. 10.1109/cvpr.2018.00716 |
17 | JEON Y, KIM J. Active convolution: learning the shape of convolution for image classification [C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 4201-4209. 10.1109/cvpr.2017.200 |
18 | LI X, WANG W H, HU X L, et al. Selective kernel networks [C]// Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2019: 510-519. 10.1109/cvpr.2019.00060 |
19 | TARMOUN S, FRANCA G, HAEFFELE B D, et al. Implicit acceleration of gradient flow in overparameterized linear models [EB/OL]. (2021-03-06) [2021-08-08]. . |
20 | CAO J M, LI Y Y, SUN M C, et al. DO-Conv: depthwise over-parameterized convolutional layer[J]. IEEE Transactions on Image Processing, 2022, 31: 3726-3736. 10.1109/tip.2022.3175432 |
21 | BOSMA W, CANNON J, PLAYOUST C. The Magma algebra system I: the user language [J]. Journal of Symbolic Computation, 1997, 24(3/4): 235-265. 10.1006/jsco.1996.0125 |
22 | HE K M, ZHANG X Y, REN S Q, et al. Delving deep into rectifiers: surpassing human-level performance on ImageNet classification [C]// Proceedings of the 2015 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2015: 1026-1034. 10.1109/iccv.2015.123 |
23 | SAXE A M, MCCLELLAND J L, GANGULI S. Exact solutions to the nonlinear dynamics of learning in deep linear neural networks [EB/OL]. (2014-02-19) [2021-09-11]. . 10.1073/pnas.1820226116 |
[1] | Dongwei WANG, Baichen LIU, Zhi HAN, Yanmei WANG, Yandong TANG. Deep network compression method based on low-rank decomposition and vector quantization [J]. Journal of Computer Applications, 2024, 44(7): 1987-1994. |
[2] | Feiyu ZHAI, Handa MA. Hybrid classical-quantum classification model based on DenseNet [J]. Journal of Computer Applications, 2024, 44(6): 1905-1910. |
[3] | Bin XIAO, Mo YANG, Min WANG, Guangyuan QIN, Huan LI. Domain generalization method of phase-frequency fusion from independent perspective [J]. Journal of Computer Applications, 2024, 44(4): 1002-1009. |
[4] | Xue LI, Guangle YAO, Honghui WANG, Jun LI, Haoran ZHOU, Shaoze YE. Remote sensing image classification based on sample incremental learning [J]. Journal of Computer Applications, 2024, 44(3): 732-736. |
[5] | Li XIE, Weiping SHU, Junjie GENG, Qiong WANG, Hailin YANG. Few-shot cervical cell classification combining weighted prototype and adaptive tensor subspace [J]. Journal of Computer Applications, 2024, 44(10): 3200-3208. |
[6] | Wen ZHOU, Yuzhang CHEN, Zhiyuan WEN, Shiqi WANG. Fish image classification based on positional overlapping patch embedding and multi-scale channel interactive attention [J]. Journal of Computer Applications, 2024, 44(10): 3209-3216. |
[7] | Tong CHEN, Jiwei WEI, Shiyuan HE, Jingkuan SONG, Yang YANG. Adversarial training method with adaptive attack strength [J]. Journal of Computer Applications, 2024, 44(1): 94-100. |
[8] | Kejun JIN, Hongtao YU, Yiteng WU, Shaomei LI, Jianpeng ZHANG, Honghao ZHENG. Improved defense method for graph convolutional network based on singular value decomposition [J]. Journal of Computer Applications, 2023, 43(5): 1511-1517. |
[9] | Bin WANG, Tian XIANG, Yidong LYU, Xiaofan WANG. Adaptive multi-scale feature channel grouping optimization algorithm based on NSGA‑Ⅱ [J]. Journal of Computer Applications, 2023, 43(5): 1401-1408. |
[10] | Kai WEN, Xiao XUE, Juan JI. Shared transformation matrix capsule network for complex image classification [J]. Journal of Computer Applications, 2023, 43(11): 3411-3417. |
[11] | Jiaxuan WEI, Shikang DU, Zhixuan YU, Ruisheng ZHANG. Review of white-box adversarial attack technologies in image classification [J]. Journal of Computer Applications, 2022, 42(9): 2732-2741. |
[12] | Wei REN, Hexiang BAI. Multi-label image classification method based on global and local label relationship [J]. Journal of Computer Applications, 2022, 42(5): 1383-1390. |
[13] | Mo LI, Tianliang LU, Ziheng XIE. Android malware family classification method based on code image integration [J]. Journal of Computer Applications, 2022, 42(5): 1490-1499. |
[14] | Yifei WANG, Lei YU, Fei TENG, Jiayu SONG, Yue YUAN. Resource load prediction model based on long-short time series feature fusion [J]. Journal of Computer Applications, 2022, 42(5): 1508-1515. |
[15] | Changqing JI, Zhiyong GAO, Jing QIN, Zumin WANG. Review of image classification algorithms based on convolutional neural network [J]. Journal of Computer Applications, 2022, 42(4): 1044-1049. |
Viewed | ||||||
Full text |
|
|||||
Abstract |
|
|||||