深度神經網絡對於訓練和測試數據分布一致的情況下取得了顯著的成功。然而，在實際情況下，測試數據多時候與訓練過程中使用的數據非常不同。現有方法通過增加訓練數據的多樣性來解決這個問題，達到涵蓋潛在的測試空間的整個範圍。儘管通常有效，但這種方法存在一些重要的限制，其中最明顯的是不能保證在視覺上不相似的樣本上有改進，並且在添加多種增強方法幾乎不會有任顯著的效果，甚至有可能降低模型的性能。為了解決這個問題，本研究通過在輸入樣本中弱化紋理且增加一個能夠集中於形狀特徵的分支來增強CNN模型的穩健性，使紋理平滑並且形狀仍然保留，以抵消受損樣本的影響。此外，本論文提出了一種形狀強化技術，進一步加強了韌性，強調形狀分支的特徵被用作標準CNN分支的特徵的指南。所提出的方法，即CGM（Cam Guided Model），在幾個開源數據集上進行了評估。結果證實，相對於現有的最先進的Augmix方法，CGM在各種常見的損壞性能指標的提升方面取得了顯著的改進。本研究提出的這種方法對於推動深度神經網絡在應對現實世界數據變化方面的穩定性和可靠性作出了有意義的貢獻。

Deep neural networks achieve remarkable success in scenarios where the distributions of the training and testing data align. However, this is seldom the case for real-world situations, in which the testing data may be very different from those used in the training process. Existing methods tackle this problem by augmenting the training data diversity, aiming to encompass the entirety of the potential testing space. While usually effective, such method suffer several important limitations, including most notably there is no guarantee improvement on samples visually dissimilar samples, and addition of multiple augmentation methods would have little to no improvement at all, or even degrade the performance of the model. To address this issue, the present study enhances the robustness of CNN models through the integration of a shape-enforcing branch in the texture of the input samples is smoothed and the shape is still preserved, to counteract the impact of corrupted samples. Moreover, a shape-enforcing technique is proposed to further fortify the resilience, the features of the shape-enforcing branch are used as a guide for features of the standard vanilla CNN branch. The efficacy of the proposed method, designated as CGM, Cam Guided Model is assessed across several open-source datasets. The results confirm that CGM achieved a marked improvement over existing state-of-the-art Augmix, as evidenced by the enhancement of various common corruption performance metrics. This method proposed in this study makes a meaningful contribution to advancing the stability and reliability of deep neural networks in confronting real-world data variations.

[1] P. Benz, C. Zhang, A. Karjauv, and I. S. Kweon, “Revisiting Batch Normalization for Improving Corruption Robustness,” in 2021 IEEE Winter Conference on Applications of Computer Vision (WACV), Waikoloa, HI, USA: IEEE, Jan. 2021, pp. 494–503. doi: 10.1109/WACV48630.2021.00054.
[2] S. Bhojanapalli, A. Chakrabarti, D. Glasner, D. Li, T. Unterthiner, and A. Veit, “Understanding Robustness of Transformers for Image Classification,” in 2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada: IEEE, Oct. 2021, pp. 10211–10221. doi: 10.1109/ICCV48922.2021.01007.
[3] G. Chen, P. Peng, L. Ma, J. Li, L. Du, and Y. Tian, “Amplitude-Phase Recombination: Rethinking Robustness of Convolutional Neural Networks in Frequency Domain,” in 2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada: IEEE, Oct. 2021, pp. 448–457. doi: 10.1109/ICCV48922.2021.00051.
[4] I. Çugu, M. Mancini, Y. Chen, and Z. Akata, “Attention Consistency on Visual Corruptions for Single-Source Domain Generalization,” in IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, CVPR Workshops 2022, New Orleans, LA, USA, June 19-20, 2022, IEEE, 2022, pp. 4164–4173. doi: 10.1109/CVPRW56347.2022.00461.
[5] T. Devries and G. W. Taylor, “Improved Regularization of Convolutional Neural Networks with Cutout,” CoRR, vol. abs/1708.04552, 2017, Accessed: Aug. 12, 2023. [Online]. Available: http://arxiv.org/abs/1708.04552
[6] N. B. Erichson, S. H. Lim, F. Utrera, W. Xu, Z. Cao, and M. W. Mahoney, “NoisyMix: Boosting Robustness by Combining Data Augmentations, Stability Training, and Noise Injections,” CoRR, vol. abs/2202.01263, 2022, [Online]. Available: https://arxiv.org/abs/2202.01263
[7] R. Geirhos, P. Rubisch, C. Michaelis, M. Bethge, F. A. Wichmann, and W. Brendel, “ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness,” in 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019, OpenReview.net, 2019. [Online]. Available: https://openreview.net/forum?id=Bygh9j09KX
[8] I. J. Goodfellow, J. Shlens, and C. Szegedy, “Explaining and Harnessing Adversarial Examples,” in 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, Y. Bengio and Y. LeCun, Eds., 2015. Accessed: Aug. 12, 2023. [Online]. Available: http://arxiv.org/abs/1412.6572
[9] K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” in 2016 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, Las Vegas, NV, USA, June 27-30, 2016, IEEE Computer Society, 2016, pp. 770–778. doi: 10.1109/CVPR.2016.90.
[10] D. Hendrycks et al., “The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization,” in 2021 IEEE/CVF International Conference on Computer Vision, ICCV 2021, Montreal, QC, Canada, October 10-17, 2021, IEEE, 2021, pp. 8320–8329. doi: 10.1109/ICCV48922.2021.00823.
[11] D. Hendrycks and T. G. Dietterich, “Benchmarking Neural Network Robustness to Common Corruptions and Perturbations,” in 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019, OpenReview.net, 2019. [Online]. Available: https://openreview.net/forum?id=HJz6tiCqYm
[12] D. Hendrycks, K. Lee, and M. Mazeika, “Using Pre-Training Can Improve Model Robustness and Uncertainty,” in Proceedings of the 36th International Conference on Machine Learning, ICML 2019, 9-15 June 2019, Long Beach, California, USA, K. Chaudhuri and R. Salakhutdinov, Eds., in Proceedings of Machine Learning Research, vol. 97. PMLR, 2019, pp. 2712–2721. [Online]. Available: http://proceedings.mlr.press/v97/hendrycks19a.html
[13] D. Hendrycks, N. Mu, E. D. Cubuk, B. Zoph, J. Gilmer, and B. Lakshminarayanan, “AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty,” in 8th International Conference on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020, OpenReview.net, 2020. [Online]. Available: https://openreview.net/forum?id=S1gmrxHFvB
[14] G. Huang, Z. Liu, L. van der Maaten, and K. Q. Weinberger, “Densely Connected Convolutional Networks,” in 2017 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, Honolulu, HI, USA, July 21-26, 2017, IEEE Computer Society, 2017, pp. 2261–2269. doi: 10.1109/CVPR.2017.243.
[15] A. Madry, A. Makelov, L. Schmidt, D. Tsipras, and A. Vladu, “Towards Deep Learning Models Resistant to Adversarial Attacks,” in 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, Conference Track Proceedings, OpenReview.net, 2018. [Online]. Available: https://openreview.net/forum?id=rJzIBfZAb
[16] E. Mintun, A. Kirillov, and S. Xie, “On Interaction Between Augmentations and Corruptions in Natural Corruption Robustness,” in Advances in Neural Information Processing Systems 34: Annual Conference on Neural Information Processing Systems 2021, NeurIPS 2021, December 6-14, 2021, virtual, M. Ranzato, A. Beygelzimer, Y. N. Dauphin, P. Liang, and J. W. Vaughan, Eds., 2021, pp. 3571–3583. [Online]. Available: https://proceedings.neurips.cc/paper/2021/hash/1d49780520898fe37f0cd6b41c5311bf-Abstract.html
[17] A. Modas, R. Rade, G. Ortiz-Jiménez, S.-M. Moosavi-Dezfooli, and P. Frossard, “PRIME: A few primitives can boost robustness to common corruptions.” arXiv, Mar. 13, 2022. doi: 10.48550/arXiv.2112.13547.
[18] T. Saikia, C. Schmid, and T. Brox, “Improving robustness against common corruptions with frequency biased models,” in 2021 IEEE/CVF International Conference on Computer Vision, ICCV 2021, Montreal, QC, Canada, October 10-17, 2021, IEEE, 2021, pp. 10191–10200. doi: 10.1109/ICCV48922.2021.01005.
[19] M. Tan and Q. V. Le, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” in Proceedings of the 36th International Conference on Machine Learning, ICML 2019, 9-15 June 2019, Long Beach, California, USA, K. Chaudhuri and R. Salakhutdinov, Eds., in Proceedings of Machine Learning Research, vol. 97. PMLR, 2019, pp. 6105–6114. [Online]. Available: http://proceedings.mlr.press/v97/tan19a.html
[20] I. Vasiljevic, A. Chakrabarti, and G. Shakhnarovich, “Examining the Impact of Blur on Recognition by Convolutional Networks,” CoRR, vol. abs/1611.05760, 2016, [Online]. Available: http://arxiv.org/abs/1611.05760
[21] H. Wang, X. Wu, Z. Huang, and E. P. Xing, “High-Frequency Component Helps Explain the Generalization of Convolutional Neural Networks,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2020, Seattle, WA, USA, June 13-19, 2020, Computer Vision Foundation / IEEE, 2020, pp. 8681–8691. doi: 10.1109/CVPR42600.2020.00871.
[22] Q. Xie, M.-T. Luong, E. H. Hovy, and Q. V. Le, “Self-Training With Noisy Student Improves ImageNet Classification,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2020, Seattle, WA, USA, June 13-19, 2020, Computer Vision Foundation / IEEE, 2020, pp. 10684–10695. doi: 10.1109/CVPR42600.2020.01070.
[23] S. Xie, R. B. Girshick, P. Dollár, Z. Tu, and K. He, “Aggregated Residual Transformations for Deep Neural Networks,” in 2017 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, Honolulu, HI, USA, July 21-26, 2017, IEEE Computer Society, 2017, pp. 5987–5995. doi: 10.1109/CVPR.2017.634.
[24] D. Yin, R. G. Lopes, J. Shlens, E. D. Cubuk, and J. Gilmer, “A Fourier Perspective on Model Robustness in Computer Vision,” in Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019, NeurIPS 2019, December 8-14, 2019, Vancouver, BC, Canada, H. M. Wallach, H. Larochelle, A. Beygelzimer, F. d’Alché-Buc, E. B. Fox, and R. Garnett, Eds., 2019, pp. 13255–13265. [Online]. Available: https://proceedings.neurips.cc/paper/2019/hash/b05b57f6add810d3b7490866d74c0053-Abstract.html
[25] S. Yun, D. Han, S. Chun, S. J. Oh, Y. Yoo, and J. Choe, “CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features,” in 2019 IEEE/CVF International Conference on Computer Vision, ICCV 2019, Seoul, Korea (South), October 27 - November 2, 2019, IEEE, 2019, pp. 6022–6031. doi: 10.1109/ICCV.2019.00612.
[26] H. Zhang, Y. Yu, J. Jiao, E. P. Xing, L. E. Ghaoui, and M. I. Jordan, “Theoretically Principled Trade-off between Robustness and Accuracy,” in Proceedings of the 36th International Conference on Machine Learning, ICML 2019, 9-15 June 2019, Long Beach, California, USA, K. Chaudhuri and R. Salakhutdinov, Eds., in Proceedings of Machine Learning Research, vol. 97. PMLR, 2019, pp. 7472–7482. [Online]. Available: http://proceedings.mlr.press/v97/zhang19p.html
[27] H. Zhang, M. Cissé, Y. N. Dauphin, and D. Lopez-Paz, “mixup: Beyond Empirical Risk Minimization,” in 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings, OpenReview.net, 2018. [Online]. Available: https://openreview.net/forum?id=r1Ddp1-Rb
[28] S. Zheng, Y. Song, T. Leung, and I. J. Goodfellow, “Improving the Robustness of Deep Neural Networks via Stability Training,” in 2016 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, Las Vegas, NV, USA, June 27-30, 2016, IEEE Computer Society, 2016, pp. 4480–4488. doi: 10.1109/CVPR.2016.485.
[29] B. Zhou, A. Khosla, À. Lapedriza, A. Oliva, and A. Torralba, “Learning Deep Features for Discriminative Localization,” in 2016 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, Las Vegas, NV, USA, June 27-30, 2016, IEEE Computer Society, 2016, pp. 2921–2929. doi: 10.1109/CVPR.2016.319.
[30] D. Zhou et al., “Removing Adversarial Noise in Class Activation Feature Space,” in 2021 IEEE/CVF International Conference on Computer Vision, ICCV 2021, Montreal, QC, Canada, October 10-17, 2021, IEEE, 2021, pp. 7858–7867. doi: 10.1109/ICCV48922.2021.00778.