FILTER SPECTRAL EFFICIENCY (FSE): STRUCTURED NETWORK PRUNING VIA INTER-LAYER SPECTRAL COUPLING

(Received: 12-Jul.-2026, Revised: 2-Sep.-2026 and 4-Sep.-2026 , Accepted: 4-Sep.-2026)

Authors Soumia Laroui, Hamida Seba, Karima Amrouche, Yassine Ait Ali Yahia,

Keywords #Structured pruning #Data-free pruning #Inter-layer spectral coupling #Singular value decomposition #Deep neural networks

Abstract Structured filter pruning is an effective approach for reducing the computational cost of deep neural networks while preserving efficient deployment on standard hardware. However, most existing pruning methods evaluate filters independently, overlooking the inter-layer spectral coupling between consecutive layers, while many high- performing approaches rely on activation statistics or calibration data. We propose Filter Spectral Efficiency (FSE), a data-free structured pruning framework that estimates filter importance through the spectral coupling between adjacent layers. FSE constructs singular-value-based filter descriptors, builds inter-layer coupling matrices, derives a spectral pruning criterion and allocates sparsity through a global adaptive budget. Extensive experiments on five CNN architectures for CIFAR-10 and ResNet-50 on ImageNet demonstrate that FSE consistently achieves a favorable accuracy-compression trade-off, outperforming representative structured pruning methods at comparable parameter and FLOP reductions. On CIFAR-10, FSE reduces the FLOPs of ResNet-56 by 59.8% with only a 0.07 percentage point accuracy drop and prunes GoogLeNet by 70.7% FLOPs while improving Top-1 accuracy over H Rank by 0.73 percentage points. On ImageNet, FSE improves the Top-1 accuracy of ResNet-50 by 1.14 percentage points over H Rank at a comparable FLOP reduction. These results demonstrate that inter-layer spectral coupling provides an effective and practical foundation for data-free structured network pruning.

References

[1] E. Edozie, A. N. Shuaibu, U. K. John and B. O. Sadiq, "Comprehensive Review of Recent Developments in Visual Object Detection Based on Deep Learning," Artificial Intelligence Review, vol. 58, no. 9, p. 277, 2025.

[2] M. H. Reza et al., "A Comprehensive Review of Convolutional Neural Networks: Foundations, Enhancements and Applications," Neural Computing and Applications, vol. 38, no. 4, p. 56, 2026.

[3] K. He, X. Zhang, S. Ren and J. Sun, "Deep Residual Learning for Image Recognition," Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pp. 770-778, 2016.

[4] G. Huang, Z. Liu, L. van der Maaten and K. Q. Weinberger, "Densely Connected Convolutional Networks," Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pp. 2261-2269, 2017.

[5] C. Szegedy et al., "Going Deeper with Convolutions," Proc. of the IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), pp. 1-9, Boston, USA, 2015.

[6] H. Cheng, M. Zhang and J. Q. Shi, "A Survey on Deep Neural Network Pruning: Taxonomy, Comparison, Analysis and Recommendations," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 12, pp. 10558-10578, 2024.

[7] R. Schwartz, J. Dodge, N. A. Smith and O. Etzioni, "Green ai," Commun. ACM, vol. 63, no. 12, p. 54-63, [Online], Available: https://doi.org/10.1145/3381831, Nov. 2020.

[8] J. Yang, Z. Li, Z. Feng and Y. Xie, "A Survey on Neural Network Quantization," Proc. of the 2025 6th Int. Conf. on Computer Information and Big Data Applications, (ser. CIBDA '25), Association for Computing Machinery, p. 384-394, New York, USA, 2025.

[9] S. Zhang and R. Saab, "Theoretical Guarantees for Low-rank Compression of Deep Neural Networks," Applied and Computational Harmonic Analysis, vol. 82, p. 101837, 2026.

[10] T. Shehzadi, R. Noor, I. Ifza, M. Liwicki, D. Stricker and M. Z. Afzal, "Knowledge Distillation in Object Detection: A Survey from CNN to Transformer," Sensors (Basel, Switzerland), vol. 26, no. 1, p. 292, 2026.

[11] K. Zhu, F. Hu, Y. Ding, W. Zhou and R. Wang, "A Comprehensive Review of Network Pruning Based on Pruning Granularity and Pruning Time Perspectives," Neurocomputing, vol. 626, p. 129382, [Online], Available: https://www.sciencedirect.com/science/article/pii/S09252312250005
42, 2025.

[12] S. Han, J. Pool, J. Tran and W. J. Dally, "Learning both weights and connections for efficient neural networks," in Advances in Neural Information Processing Systems (NeurIPS), 2015, pp. 1135-1143.

[13] H. Li, A. Kadav, I. Durdanovic, H. Samet and H. P. Graf, "Pruning filters for efficient convnets," in International Conference on Learning Representations (ICLR), 2017.

[14] Y. He, P. Liu, Z. Wang, Z. Hu and Y. Yang, "Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Acceleration," Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR), pp. 4340-4349, 2019.

[15] M. Lin, R. Ji, Y. Wang, Y. Zhang, B. Zhang, Y. Tian and L. Shao, "HRank: Filter Pruning Using High-rank Feature Map," Proceedings of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR), pp. 1529-1538, 2020.

[16] S. Lin et al., "Towards Optimal Structured CNN Pruning via Generative Adversarial Learning," Proc. of the 2019 IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR), pp. 2785-2794, 2019.

[17] T. Yuan, S. Yang, B. Liu and Y. Tang, "Explainable CNN Filter Pruning Based on Shapley Value Approximation in View of Game Theory," Expert Systems with Appl., vol. 300, p. 130232, 2025.

[18] E. Chen, H. Wang, Z. Shi and W. Zhang, "Channel Pruning for Convolutional Neural Networks Using l0-Norm Constraints," Neurocomputing, vol. 636, p. 129925, 2025.

[19] Ambuj, "Feature Representation Fidelity Preservation During Neural Network Pruning for Enhanced Compression Efficiency," Neurocomputing, vol. 634, p. 129901, 2025.

[20] A. Krizhevsky, "Learning Multiple Layers of Features from Tiny Images," University of Toronto, Technical Report, [Online], Available: https://cave.cs.toronto.edu/kriz/learning-features-2009-TR.pdf, 2009.

[21] K. Simonyan and A. Zisserman, "Very Deep Convolutional Networks for Large-scale Image Recognition," Proc. of the Int. Conf. on Learning Representations (ICLR), 2015.

[22] O. Russakovsky et al., "ImageNet Large Scale Visual Recognition Challenge," Int. Journal of Computer Vision, vol. 115, no. 3, pp. 211-252, 2015.

[23] Y. Sui et al., "Chip: Channel Independence-based Pruning for Compact Neural Networks," Advances in Neural Information Processing Systems, vol. 34, pp. 24604-24616, 2021.

[24] M. Lin, L. Cao, S. Li, Q. Ye, Y. Tian, J. Liu, Q. Tian and R. Ji, "Filter Sketch for Network Pruning," IEEE Transactions on Neural Networks and Learning Systems, vol. 33, no. 12, pp. 7091-7100, 2021.

[25] J.-H. Luo et al., "ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression," Proc. of the IEEE Int. Conf. on Computer Vision, pp. 5058-5066, Venice, Italy, 2017.

[26] Y. He, X. Zhang and J. Sun, "Channel Pruning for Accelerating Very Deep Neural Networks," Proc. of the IEEE Int. Conf. on Computer Vision, pp. 1389-1397, Venice, Italy, 2017.

[27] M. Alwani, Y. Wang and V. Madhavan, "DECORE: Deep Compression with Reinforcement Learning," Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition, pp. 12349-12359, New Orleans, USA, 2022.

[28] Y. Zheng, P. Sun, Q. Ren, W. Xu and D. Zhu, "A Novel and Efficient Model Pruning Method for Deep Convolutional Neural Networks by Evaluating the Direct and Indirect Effects of Filters," Neurocomputing, vol. 569, p. 127124, 2024.

[29] Y. Lian, P. Peng, K. Jiang and W. Xu, "Cross-layer Importance Evaluation for Neural Network Pruning," Neural Networks, vol. 179, p. 106496, 2024.

[30] Y. He et al., "Soft Filter Pruning for Accelerating Deep Convolutional Neural Networks," Proc. of the 27th Int. Joint Conf. on Artificial Intelligence (IJCAI), pp. 2234-2240, 2018.

[31] V. T. Pham, Y. Zniyed and T. P. Nguyen, "Efficient Tensor Decomposition-based Filter Pruning," Neural Networks, vol. 178, p. 106393, 2024.

[32] C. Zhen, W. Zhang, J. Mo, M. Ji, H. Zhou and J. Zhu, "Rasp: Regularization-based Amplitude Saliency Pruning," Neural Networks, vol. 168, pp. 1-13, 2023.

[33] Z. Liu, J. Li, Z. Shen, G. Huang, S. Yan and C. Zhang, "Learning Efficient Convolutional Networks through Network Slimming," Proc. of the 2017 IEEE Int. Conf. on Computer Vision (ICCV), pp. 2755-2763, Venice, Italy, 2017.

[34] M. Lin, R. Ji, Y. Zhang, B. Zhang, Y. Wu and Y. Tian, "Channel Pruning via Automatic Structure Search," Proc. of the 29th Int. Joint Conf. on Artificial Intelligence (IJCAI'20), pp. 673-679, 2021.

[35] Y. Xue, W. Yao, S. Peng and S. Yao, "Automatic Filter Pruning Algorithm for Image Classification," Applied Intelligence, vol. 54, no. 1, pp. 216-230, 2024.

[36] V. T. Pham et al., "Enhanced Network Compression through Tensor Decompositions and Pruning," IEEE Trans. on Neural Networks and Learning Systems, vol. 36, no. 3, pp. 4358-4370, 2025.

[37] V. T. Pham et al., "Coupled Tensor Decomposition for Compact Network Representation," IEEE Trans. on Neural Networks and Learning Systems, vol. 37, no. 2, pp. 617-631, 2026.

[38] W. Lin et al., "STPP: Efficient and Progressive Structured Pruning via Enhanced Sparsification Paradigm," IEEE Trans. on Pattern Analysis and Machine Intelligence, Early Publishing, pp. 1-17, 2026.

[39] G. Fang, X. Ma, M. Song, M. Bi Mi and X. Wang, "DepGraph: Towards Any Structural Pruning," Proc. of the 2023 IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR), pp. 16091-16 101, Vancouver, Canada, 2023.

[40] Z. Huang and N. Wang, "Data-driven Sparse Structure Selection for Deep Neural Networks," Proc. of the 15th European Conf. on Computer Vision (ECCV 2018), Part XVI, pp. 317-334, Munich, Germany, September 8-14, 2018.

[41] A. Schrijver, Combinatorial Optimization: Polyhedra and Efficiency, E-ISSN: 2197-6783, Springer, 2003.