CDTNet: Improved Image Classification Method Using Standard, Dilated and Transposed Convolutions
Source abstract
Abstract
Convolutional neural networks (CNNs) have achieved great success in image classification tasks. In the process of a convolutional operation, a larger input area can capture more context information. Stacking several convolutional layers can enlarge the receptive field, but this increases the parameters. Most CNN models use pooling layers to extract important features, but the pooling operations cause information loss. Transposed convolution can increase the spatial size of the feature maps to recover the lost low-resolution information. In this study, we used two branches with different dilated rates to obtain different size features. The dilated convolution can capture richer information, and the outputs from the two channels are concatenated together as input for the next block. The small size feature maps of the top blocks are transposed to increase the spatial size of the feature maps to recover low-resolution prediction maps. We evaluated the model on three image classification benchmark datasets (CIFAR-10, SVHN, and FMNIST) with four state-of-the-art models, namely, VGG16, VGG19, ResNeXt, and DenseNet. The experimental results show that CDTNet achieved lower loss, higher accuracy, and faster convergence speed in the training and test stages. The average test accuracy of CDTNet increased by 54.81% at most on SVHN with VGG19 and by 1.28% at least on FMNIST with VGG16, which proves that CDTNet has better performance and strong generalization abilities, as well as fewer parameters.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Advanced Neural Network Applications · Domain Adaptation and Few-Shot Learning · COVID-19 diagnosis using AI
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou 510006, China
School of Information Management, Sun Yat-sen University, Guangzhou 510006, China
School of Information and Engineering, Shaoguan University, Shaoguan 512005, China
Funding
Basic and Applied Basic Research Fund of Guangdong Province
References (76)
Show source-provided references
1. Gradient-based learning applied to document recognition10.1109/5.726791 ↗
2. Krizhevsky, A., Sutskever, I., and Hinton, G.E. (2012, January 3–6). Imagenet classification with deep convolutional neural networks. Proceedings of the Advances in Neural Information Processing Systems, Carson, NV, USA.
3. Simonyan, K., and Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv.
4. Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., and Rabinovich, A. (2015, January 7–12). Going deeper with convolutions. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA.10.1109/CVPR.2015.7298594 ↗
5. Local Deep-Feature Alignment for Unsupervised Dimension Reduction10.1109/TIP.2018.2804218 ↗
6. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks10.1109/TPAMI.2016.2577031 ↗
7. Kim, Y. (2014, January 25–29). Convolutional Neural Networks for Sentence Classification. Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP 2014), Doha, Qatar.10.3115/v1/D14-1181 ↗
8. Lin, T.Y., Dollár, P., Girshick, R., He, K., Hariharan, B., and Belongie, S. (2017, January 21–26). Feature pyramid networks for object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA.10.1109/CVPR.2017.106 ↗
9. Huang, G., Liu, Z., Van Der Maaten, L., and Weinberger, K.Q. (2017, January 21–26). Densely connected convolutional networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA.10.1109/CVPR.2017.243 ↗
10. Tian, Z., Shen, C., Chen, H., and He, T. (2019, January 27–28). FCOS: Fully Convolutional One-Stage Object Detection. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, Korea.10.1109/ICCV.2019.00972 ↗
11. Yilmazer, R., and Birant, D. (2021). Shelf Auditing Based on Image Classification Using Semi-Supervised Deep Learning to Increase On-Shelf Availability in Grocery Stores. Sensors, 21.10.3390/s21020327 ↗
12. Zeng, J., Zhang, D., Li, Z., and Li, X. (2021). Semi-Supervised Training of Transformer and Causal Dilated Convolution Network with Applications to Speech Topic Classification. Appl. Sci., 11.10.3390/app11125712 ↗
13. Automatic Calcium Scoring in Low-Dose Chest CT Using Deep Neural Networks with Dilated Convolutions10.1109/TMI.2017.2769839 ↗
14. Md-Net: Multi-scale Dilated Convolution Network for CT Images Segmentation10.1007/s11063-020-10230-x ↗
15. Wang, T., Sun, M., and Hu, K. (2017, January 6–8). Dilated Deep Residual Network for Image Denoising. Proceedings of the IEEE 29th International Conference on Tools with Artificial Intelligence (ICTAI), Boston, MA, USA.10.1109/ICTAI.2017.00192 ↗
16. Attention-guided CNN for image denoising10.1016/j.neunet.2019.12.024 ↗
17. Dilated Residual Networks with Symmetric Skip Connection for image denoising10.1016/j.neucom.2018.12.075 ↗
18. Fully Convolutional Networks for Semantic Segmentation10.1109/TPAMI.2016.2572683 ↗
19. Peng, C., Zhang, X., Yu, G., Luo, G., and Sun, J. (2017, January 21–26). Large Kernel Matters-Improve Semantic Segmentation by Global Convolutional Network. Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA.10.1109/CVPR.2017.189 ↗
20. DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs10.1109/TPAMI.2017.2699184 ↗
21. Improving the Performance of VGG Through Different Granularity Feature Combinations10.1109/ACCESS.2020.3031908 ↗
22. Structure optimization of convolutional neural networks: A survey
23. Compensating Delays and Noises in Motion Control of Autonomous Electric Vehicles by Using Deep Learning and Unscented Kalman Predictor10.1109/TSMC.2018.2850367 ↗
24. Looking Closer at the Scene: Multiscale Representation Learning for Remote Sensing Image Scene Classification10.1109/TNNLS.2020.3042276 ↗
25. He, K., Zhang, X., Ren, S., and Sun, J. (2016, January 27–30). Deep residual learning for image recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA.10.1109/CVPR.2016.90 ↗
26. Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., and Wojna, Z. (2016, January 27–30). Rethinking the Inception Architecture for Computer Vision. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA.10.1109/CVPR.2016.308 ↗
27. Receptive Field Size Versus Model Depth for Single Image Super-Resolution10.1109/TIP.2019.2941327 ↗
28. A multiscale dilated dense convolutional network for saliency prediction with instance-level attention competition10.1016/j.jvcir.2019.102611 ↗
29. Luo, W., Li, Y., Urtasun, R., and Zemel, R. (2016, January 5–10). Understanding the effective receptive field in deep convolutional neural networks. Proceedings of the 30th International Conference on Neural Information Processing Systems, Barcelona, Spain.
30. Huang, G., Liu, S., Maaten, L.V.D., and Weinberger, K.Q. (2018, January 18–23). CondenseNet: An efficient DenseNet using learned group convolutions. Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.10.1109/CVPR.2018.00291 ↗
31. Liu, Z., Sun, M., Zhou, T., Huang, G., and Darrell, T. (2018). Rethinking the value of network pruning. arXiv.
32. PAC-Bayesian framework based drop-path method for 2D discriminative convolutional network pruning10.1007/s11045-019-00686-z ↗
33. Kendall, A., and Gal, Y. (2017, January 4–9). What uncertainties do we need in bayesian deep learning for computer vision?. Proceedings of the Advances in Neural Information Processing Systems (NIPS), Long Beach, CA, USA.
34. Improving neural networks by preventing co-adaptation of feature detectors
35. Spectrum interference-based two-level data augmentation method in deep learning for automatic modulation classification10.1007/s00521-020-05514-1 ↗
36. Larsson, G., Maire, M., and Shakhnarovich, G. (2016). Fractalnet: Ultra-deep neural networks without residuals. arXiv.
37. Differential learning: A powerful tool for interactive content-based image retrieval
38. Kobayashi, T. (2017, January 22–29). Flip-invariant motion representation. Proceedings of the Proceedings of the IEEE International Conference on Computer Vision (ICCV), Venice, Italy.10.1109/ICCV.2017.600 ↗
39. A full stage data augmentation method in deep convolutional neural network for natural image classification10.1155/2020/4706576 ↗
40. Roth, V., and Vetter, T. (2017). Learning Dilation Factors for Semantic Segmentation of Street Scenes. German Conference on Pattern Recognition, Springer.10.1007/978-3-319-66709-6 ↗
41. Dilated Convolution and Feature Fusion SSD Network for Small Object Detection in Remote Sensing Images10.1109/ACCESS.2020.2991439 ↗
42. Heo, W.-H., Kim, H., and Kwon, O.-W. (2020). Source Separation Using Dilated Time-Frequency DenseNet for Music Identification in Broadcast Contents. Appl. Sci., 10.10.3390/app10051727 ↗
43. Heo, W.-H., Kim, H., and Kwon, O.-W. (2021). Integrating Dilated Convolution into DenseLSTM for Audio Source Separation. Appl. Sci., 11.10.3390/app11020789 ↗
44. Fritzsche, K., Deserno, G., Lehmann, T., Handels, H., and Tolxdorff, T. (2017). Invited Talk: U-Net: Convolutional Networks for Biomedical Image Segmentation. Bildverarbeitung für die Medizin, Springer.
45. Zeiler, M.D., Taylor, G.W., and Fergus, R. (2011, January 6–13). Adaptive deconvolutional networks for mid and high level feature learning. Proceedings of the 2011 International Conference on Computer Vision, Barcelona, Spain.10.1109/ICCV.2011.6126474 ↗
46. Xie, S., Girshick, R., Dollar, P., Tu, Z., and He, K. (2017, January 21–26). Aggregated Residual Transformations for Deep Neural Networks. Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA.10.1109/CVPR.2017.634 ↗
47. Yu, F., and Koltun, V. (2016). Multi-scale context aggregation by dilated convolutions. arXiv.
48. A hybrid convolution network for serial number recognition on banknotes10.1016/j.ins.2019.09.070 ↗
49. The classification of gliomas based on a Pyramid dilated convolution resnet model10.1016/j.patrec.2020.03.007 ↗
50. Yao, S., Chen, Y., Tian, X., Jiang, R., and Ma, S. (2020). An Improved Algorithm for Detecting Pneumonia Based on YOLOv3. Appl. Sci., 10.10.3390/app10051818 ↗
51. Cascaded hierarchical atrous spatial pyramid pooling module for semantic segmentation10.1016/j.patcog.2020.107622 ↗
52. Zeiler, M.D., Krishnan, D., Taylor, G.W., and Fergus, R. (2010, January 13–18). Deconvolutional networks. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Francisco, CA, USA.10.1109/CVPR.2010.5539957 ↗
53. Gulrajani, I., Kumar, K., Ahmed, F., Taiga, A.A., Visin, F., Vazquez, D., and Courville, A. (2016). Pixelvae: A latent variable model for natural images. arXiv.
54. A deep generative deconvolutional image model
55. Dumoulin, V., and Visin, F. (2016). A guide to convolution arithmetic for deep learning. arXiv.
56. Fuzzy license plate restoration method based on convolution and transposed convolution
57. Bukka, S.R., Gupta, R., and Magee, A.R. (2020). Assessment of unsteady flow predictions using hybrid deep learning based reduced order models. arXiv.10.1063/5.0030137 ↗
58. Contextual Deconvolution Network for Semantic Segmentation10.1016/j.patcog.2019.107152 ↗
59. Noh, H., Hong, S., and Han, B. (2015, January 7–13). Learning deconvolution network for semantic segmentation. Proceedings of the 2015 IEEE International Conference on Computer Vision (ICCV), Santiago, Chile.10.1109/ICCV.2015.178 ↗
60. Cui, Z., Chang, H., Shan, S., Zhong, B., and Chen, X. (2014, January 6–12). Deep network cascade for image super-resolution. Proceedings of the European Conference on Computer Vision, Zurich, Switzerland.10.1007/978-3-319-10602-1_4 ↗
61. Image super-resolution using a dilated convolutional neural network10.1016/j.neucom.2017.09.062 ↗
62. Anisotropic Convolution for Image Classification10.1109/TIP.2020.2985875 ↗
63. Fu, J., Liu, J., and Wang, Y. (2019). Stacked deconvolutional network for semantic segmentation. IEEE Trans. Image Process., 1–13.10.1109/TIP.2019.2895460 ↗
64. Bownet: Dilated convolution neural network for ultrasound tongue contour extraction
65. Chen, H., Sun, K., and Tian, Z. (2020, January 13–19). BlendMask: Top-Down Meets Bottom-Up for Instance Segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Seattle, WA, USA.10.1109/CVPR42600.2020.00860 ↗
66. Squeeze-and-Excitation Networks10.1109/TPAMI.2019.2913372 ↗
67. DCSR: Dilated Convolutions for Single Image Super-Resolution10.1109/TIP.2018.2877483 ↗
68. Compressed sensing MRI via a multi-scale dilated residual convolution network10.1016/j.mri.2019.07.014 ↗
69. Yu, F., Wang, D., Shelhamer, E., and Darrell, T. (2018, January 18–23). Deep Layer Aggregation. Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.10.1109/CVPR.2018.00255 ↗
70. Ioffe, S., and Szegedy, C. (2015, January 7–9). Batch normalization: Accelerating deep network training by reducing internal covariate shift. Proceedings of the 32nd International Conference on Machine Learning, Lille, France.
71. Nair, V., and Hinton, G. (2010, January 21–24). Rectified linear units improve restricted Boltzmann machines. Proceedings of the 27th International Conference on International Conference on Machine Learning, Haifa, Israel.
72. Krizhevsky, A., and Hinton, G. (2009). Learning Multiple Layers of Features from Tiny Images, University of Toronto. Technical Report.
73. Netzer, Y., Wang, T., Coates, A., Bissacco, A., Wu, B., and Ng, A.Y. (2011, January 12–15). Reading digits in natural images with unsupervised feature learning. Proceedings of the Conference on Neural Information Processing Systems (NIPS), Granada, Spain.
74. Xiao, H., Rasul, K., and Vollgraf, R. (2017). Fashion-MNIST: A Novel Image Dataset for Benchmarking Machine Learning Algorithms. arXiv.
75. Introspective analysis of convolutional neural networks for improving discrimination performance and feature visualisation10.7717/peerj-cs.497 ↗
76. Li, X., Li, F., Fern, X., and Raich, R. (2017, January 24–26). Filter shaping for convolutional neural networks. Proceedings of the ICLR 2017 Conference, Toulon, France.