WSG-FEC: Gray-Box Frequency-Domain Explanation of CNN Models Based on Weakly Supervised Learning
Source abstract
Abstract
Convolutional neural networks (CNNs) have achieved remarkable progress in recent years, largely driven by advances in computational hardware. However, their increasingly complex architectures continue to pose significant challenges for interpretability. Existing explanation methods predominantly rely on spatial saliency representations and therefore fail to capture the intrinsic frequency-domain characteristics of image data. This work introduces WSG-FEC, a Weakly Supervised Gray-box Frequency-domain Explanation framework that integrates multilevel Haar wavelet decomposition with gradient-weighted class activation mapping (Grad-CAM) to generate frequency-aware explanations. By leveraging hierarchical wavelet structures, the proposed method produces frequential–spatial saliency maps that overcome the low-resolution limitations of white-box approaches and the high computational cost of black-box perturbation methods. Quantitative evaluations using insertion and deletion metrics demonstrate that WSG-FEC provides more detailed, efficient, and interpretable explanations, offering a novel perspective for understanding CNN decision mechanisms.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Explainable Artificial Intelligence (XAI) · Advanced Neural Network Applications · Generative Adversarial Networks and Image Synthesis
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Faculty of Innovation Engineering, Macau University of Science and Technology, Macau 999078, China
References (32)
Show source-provided references
1. Bonalumi, L., Aymerich, E., Alessi, E., Cannas, B., Fanni, A., Lazzaro, E., Nowak, S., Pisano, F., Sias, G., and Sozzi, C. (2024). eXplainable Artificial Intelligence applied to algorithms for disruptions prediction in tokamak devices. Front. Phys., 12.10.3389/fphy.2024.1359656 ↗
2. Explainable deep learning for the analysis of MHD spectrograms in nuclear fusion10.1088/2632-2153/ac44aa ↗
3. Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., and Torralba, A. (2016, January 27–30). Learning Deep Features for Discriminative Localization. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2016), Las Vegas, NV, USA.10.1109/CVPR.2016.319 ↗
4. Selvaraju, R.R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., and Batra, D. (2017, January 22–29). Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization. Proceedings of the IEEE International Conference on Computer Vision (ICCV), Venice, Italy.10.1109/ICCV.2017.74 ↗
5. Chattopadhay, A., Sarkar, A., Howlader, P., and Balasubramanian, V. (2018, January 12–15). Grad-CAM++: Generalized Gradient-Based Visual Explanations for Deep Convolutional Networks. Proceedings of the IEEE Winter Conference on Applications of Computer Vision (WACV), Lake Tahoe, NV, USA.10.1109/WACV.2018.00097 ↗
6. Chen, C., Li, O., Tao, D., Barnett, A., Rudin, C., and Su, J. (2019, January 8–14). This Looks Like That: Deep Learning for Interpretable Image Recognition. Proceedings of the 33rd Conference on Neural Information Processing Systems (NeurIPS), Vancouver, BC, Canada.
7. Stefenon, S.F., Singh, G., Yow, K.C., and Cimatti, A. (2022). Semi-ProtoPNet Deep Neural Network for the Classification of Defective Power Grid Distribution Structures. Sensors, 22.10.3390/s22134859 ↗
8. Singh, G., and Yow, K.C. (2021). Object or Background: An Interpretable Deep Learning Model for COVID-19 Detection from CT-Scan Images. Diagnostics, 11.10.3390/diagnostics11091732 ↗
9. One and One Make Eleven: An Interpretable Neural Network for Image Recognition10.1016/j.knosys.2023.110926 ↗
10. Petsiuk, V., Das, A., and Saenko, K. (2018). RISE: Randomized Input Sampling for Explanation of Black-Box Models. arXiv.
11. GSM-HM: Generation of Saliency Maps for Black-Box Object Detection Model Based on Hierarchical Masking10.1109/ACCESS.2022.3206379 ↗
12. Evaluating explainable artificial intelligence methods for multi-label deep learning classification tasks in remote sensing
13. Ribeiro, M.T., Singh, S., and Guestrin, C. (2016, January 13–17). “Why Should I Trust You?”: Explaining the Predictions of Any Classifier. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA.10.1145/2939672.2939778 ↗
14. Rabbani, M., and Jones, P. (1991). Digital Image Compression Techniques, SPIE Press.10.1117/3.34917 ↗
15. Xu, Z.Q.J., Zhang, Y., Luo, T., Xiao, Y., and Ma, Z. (2019). Frequency Principle: Fourier Analysis Sheds Light on Deep Neural Networks. arXiv.
16. Huber, M., and Damer, N. (March, January 26). Beyond Spatial Explanations: Explainable Face Recognition in the Frequency Domain. Proceedings of the 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Tucson, AZ, USA.10.1109/WACV61041.2025.00108 ↗
17. Huber, M., Neto, P.C., Sequeira, A.F., and Damer, N. (March, January 28). FX-MAD: Frequency-Domain Explainability and Explainability-Driven Unsupervised Detection of Face Morphing Attacks. Proceedings of the Winter Conference on Applications of Computer Vision, Tucson, AZ, USA.
18. Chliński, K., and Wawrzyński, P. (2020, January 19–24). DCT-Conv: Coding Filters in Convolutional Networks with Discrete Cosine Transform. Proceedings of the 2020 International Joint Conference on Neural Networks (IJCNN), Glasgow, UK.10.1109/IJCNN48605.2020.9207103 ↗
19. 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 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA.10.1109/CVPR.2017.106 ↗
20. Xu, Z., Zhang, Y., and Xiao, Y. (2019, January 12–15). Training Behavior of Deep Neural Network in Frequency Domain. Proceedings of the 26th International Conference on Neural Information Processing (ICONIP 2019), Part I, Sydney, NSW, Australia.10.1007/978-3-030-36708-4_22 ↗
21. U 2 D 2 Net: Unsupervised Unified Image Dehazing and Denoising Network for Single Hazy Image Enhancement10.1109/TMM.2023.3263078 ↗
22. Quantum Image Edge Detection Based on Haar Wavelet Transform10.1007/s11128-024-04513-8 ↗
23. (2000). Information Technology—JPEG 2000 Image Coding System: Core Coding System (Standard No. ISO/IEC 15444-1:2000).
24. Alsoufi, M.A., Siraj, M.M., Ghaleb, F.A., Abdulqader, A.H., Ali, E., and Omar, M. (2023). An Anomaly Intrusion Detection Systems in IoT Based on Autoencoder: A Review. Advances in Intelligent Computing Techniques and Applications. IRICT 2023. Lecture Notes on Data Engineering and Communications Technologies, Springer Nature.10.1007/978-3-031-59707-7_20 ↗
25. A Review on Autoencoder Based Representation Learning for Fault Detection and Diagnosis in Industrial Processes10.1016/j.chemolab.2022.104711 ↗
26. Hu, M., Han, H., Shan, S., and Chen, X. (2019, January 16–20). Weakly Supervised Image Classification through Noise Regularization. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, CA, USA.10.1109/CVPR.2019.01178 ↗
27. Weakly supervised machine learning10.1049/cit2.12216 ↗
28. Durand, T., Mordan, T., Thome, N., and Cord, M. (2017, January 21–26). WILDCAT: Weakly Supervised Learning of Deep ConvNets for Image Classification, Pointwise Localization and Segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA.10.1109/CVPR.2017.631 ↗
29. Fong, R.C., and Vedaldi, A. (2017, January 22–29). Interpretable Explanations of Black Boxes by Meaningful Perturbation. Proceedings of the IEEE International Conference on Computer Vision (ICCV), Venice, Italy.10.1109/ICCV.2017.371 ↗
30. Evaluating the Visualization of What a Deep Neural Network Has Learned10.1109/TNNLS.2016.2599820 ↗
31. Valois, P.H.V., Niinuma, K., and Fukui, K. (2024, January 3–8). Occlusion Sensitivity Analysis with Augmentation Subspace Perturbation in Deep Feature Space. Proceedings of the 2024 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Waikoloa, HI, USA.10.1109/WACV57701.2024.00476 ↗
32. Yang, M., and Kim, B. (2019). Benchmarking Attribution Methods with Relative Feature Importance. arXiv.