Subspace representations in deep neural networks: A survey
DOI:
https://doi.org/10.5753/jbcs.2026.5482Keywords:
Deep Learning, Subspace Methods, Manifold LearningAbstract
Computer vision applications often involve processing large-scale multidimensional data, requiring methods that are both efficient and accurate. Traditional pattern recognition methods based on subspace representations offer low computational complexity but typically underperform compared to deep learning models in terms of recognition accuracy. This study aims to explore and analyze the integration of subspace representations within deep learning frameworks to leverage the advantages of both approaches. We conducted a comprehensive survey of existing methods that combine subspace representation techniques with deep neural networks. We propose a taxonomy to categorize these methods into three distinct groups based on their integration strategies. The reviewed methods demonstrate that incorporating subspace representations can enhance the performance and efficiency of deep learning models. The taxonomy helps to clarify the landscape of these hybrid approaches and identifies trends in methodological development. The surveyed approaches demonstrate a clear methodological evolution, contributing to enhanced outcomes in various real-world applications.
Downloads
References
Belkin, M. and Niyogi, P. (2003). Laplacian eigenmaps for dimensionality reduction and data representation. Neural computation, 15(6):1373-1396. DOI: 10.1162/089976603321780317.
Bloom, V., Makris, D., and Argyriou, V. (2012). G3d: A gaming action dataset and real time action recognition evaluation framework. In 2012 IEEE Computer society conference on computer vision and pattern recognition workshops, pages 7-12. IEEE. DOI: 10.1109/cvprw.2012.6239175.
Bonnabel, S. (2013a). Stochastic gradient descent on riemannian manifolds. IEEE Transactions on Automatic Control, 58(9):2217-2229. DOI: 10.1109/tac.2013.2254619.
Bonnabel, S. (2013b). Stochastic gradient descent on riemannian manifolds. IEEE Transactions on Automatic Control, 58(9):2217-2229. DOI: 10.1109/tac.2013.2254619.
Bottou, L. (2010). Large-scale machine learning with stochastic gradient descent. In Proceedings of COMPSTAT'2010: 19th International Conference on Computational StatisticsParis France, August 22-27, 2010 Keynote, Invited and Contributed Papers, pages 177-186. Springer. DOI: 10.1201/b11429-4.
Cetingul, H. E. and Vidal, R. (2009). Intrinsic mean shift for clustering on stiefel and grassmann manifolds. In 2009 IEEE Conference on Computer Vision and Pattern Recognition, pages 1896-1902. IEEE. DOI: 10.1109/cvpr.2009.5206806.
Chen, Z., Wu, X., Xu, T., Wang, R., huang2017deep2017, Z., and Kittler, J. (2022). Msnet: A deep multi-scale submanifold network for visual classification. CoRR, abs/2201.10145. Available at:[link].
Du, Y., Wang, W., and Wang, L. (2015). Hierarchical recurrent neural network for skeleton based action recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1110-1118. DOI: 10.1109/cvpr.2015.7298714.
Edelman, A., Arias, T. A., and Smith, S. T. (1998). The geometry of algorithms with orthogonality constraints. SIAM journal on Matrix Analysis and Applications, 20(2):303-353. DOI: 10.1137/s0895479895290954.
Fei, Y., Wei, X., Liu, Y., Li, Z., and Chen, M. (2023). A survey of geometric optimization for deep learning: From euclidean space to riemannian manifold. arXiv preprint arXiv:2302.08210. DOI: 10.1145/3708498.
Gatto, B. B., dos Santos, E. M., Koerich, A. L., Fukui, K., and Junior, W. S. (2021). Tensor analysis with n-mode generalized difference subspace. Expert Systems with Applications, 171:114559. DOI: 10.1016/j.eswa.2020.114559.
Haitman, Y., Francos, J. M., and Scharf, L. L. (2021a). Grassmannian dimensionality reduction for optimized universal manifold embedding representation of 3d point clouds. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 4213-4221. DOI: 10.1109/iccvw54120.2021.00468.
Haitman, Y., Francos, J. M., and Scharf, L. L. (2021b). Grassmannian dimensionality reduction for optimized universal manifold embedding representation of 3d point clouds. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 4213-4221. DOI: 10.1109/iccvw54120.2021.00468.
Hamm, J. and Lee, D. (2008a). Extended grassmann kernels for subspace-based learning. Advances in neural information processing systems, 21. Available at:[link].
Hamm, J. and Lee, D. D. (2008b). Grassmann discriminant analysis: A unifying view on subspace-based learning. In Proceedings of the 25th International Conference on Machine Learning, ICML '08, page 376–383, New York, NY, USA. Association for Computing Machinery. DOI: 10.1145/1390156.1390204.
Harandi, M. T., Sanderson, C., Shirazi, S., and Lovell, B. C. (2011). Graph embedding discriminant analysis on grassmannian manifolds for improved image set matching. In CVPR 2011, pages 2705-2712. IEEE. DOI: 10.1109/cvpr.2011.5995564.
Hong, J., Li, Y., and Chen, H. (2019). Variant grassmann manifolds: A representation augmentation method for action recognition. ACM Transactions on Knowledge Discovery from Data (TKDD), 13(2):1-23. DOI: 10.1145/3314203.
Huang, Z., Wang, R., Shan, S., and Chen, X. (2015). Projection metric learning on grassmann manifold with application to video based face recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 140-149. DOI: 10.1109/cvpr.2015.7298609.
Ionescu, C., Vantzos, O., and Sminchisescu, C. (2015). Matrix backpropagation for deep networks with structured layers. In Proceedings of the IEEE international conference on computer vision, pages 2965-2973. DOI: 10.1109/iccv.2015.339.
Jiang, X., Hu, Z., Wang, S., and Zhang, Y. (2023). Deep learning for medical image-based cancer diagnosis. Cancers, 15(14):3608. DOI: 10.3390/cancers15143608.
Johnson, M. and Savakis, A. (2015). L1-grassmann manifolds for robust face recognition. In 2015 IEEE International Conference on Image Processing (ICIP), pages 482-486. IEEE. DOI: 10.1109/icip.2015.7350845.
Katayama, T. (2005). Subspace methods for system identification. Springer, 1 edition. DOI: 10.1007/1-84628-158-x.
Ke, Z., Cui, Z.-X., Huang, W., Cheng, J., Jia, S., Ying, L., Zhu, Y., and Liang, D. (2021). Deep manifold learning for dynamic mr imaging. IEEE Transactions on Computational Imaging, 7:1314-1327. DOI: 10.1109/tci.2021.3131564.
Kuutti, S., Bowden, R., Jin, Y., Barber, P., and Fallah, S. (2020). A survey of deep learning applications to autonomous vehicle control. IEEE Transactions on Intelligent Transportation Systems, 22(2):712-733. DOI: 10.1109/tits.2019.2962338.
Li, Y., Guo, T., Liu, X., and Xia, R. (2019). Skeleton-based action recognition with lie group and deep neural networks. In 2019 IEEE 4th International Conference on Signal and Image Processing (ICSIP), pages 26-30. IEEE. DOI: 10.1109/siprocess.2019.8868548.
Liu, S., Zhao, C., An, Y., Li, P., Zhao, J., and Zhang, Y. (2019). Diffusion tensor imaging denoising based on riemannian geometric framework and sparse bayesian learning. Journal of Medical Imaging and Health Informatics, 9(9):1993-2003. DOI: 10.1166/jmihi.2019.2832.
Liu, T., Shi, Z., Liu, Y., and Li, C. (2018). Dimensionality reduction on grassmannian: A good practice. In 2018 Eighth International Conference on Instrumentation & Measurement, Computer, Communication and Control (IMCCC), pages 943-948. IEEE. DOI: 10.1109/imccc.2018.00199.
Lu, W., Fan, F., Chu, J., Jing, P., and Yuting, S. (2018). Wearable computing for internet of things: A discriminant approach for human activity recognition. IEEE Internet of Things Journal, 6(2):2749-2759. DOI: 10.1109/jiot.2018.2873594.
Müller, M., Röder, T., Clausen, M., Eberhardt, B., Krüger, B., and Weber, A. (2007). Mocap database hdm05. Institut für Informatik II, Universität Bonn, 2(7). Available at:[link].
Nie, S. and Ji, Q. (2014). Capturing global and local dynamics for human action recognition. In 2014 22nd International Conference on Pattern Recognition, pages 1946-1951. IEEE. DOI: 10.1109/icpr.2014.340.
Oja, E. and Kut, M. (1983). The alsm algorithm - an improved subspace method of classification. Pattern Recoflnition, 16 edition. DOI: 10.1016/0031-3203(83)90064-x.
Pierson, H. A. and Gashler, M. S. (2017). Deep learning in robotics: a review of recent research. Advanced Robotics, 31(16):821-835. DOI: 10.1080/01691864.2017.1365009.
Rhif, M., Wannous, H., and Farah, I. R. (2018). Action recognition from 3d skeleton sequences using deep networks on lie group features. In 2018 24th International Conference on Pattern Recognition (ICPR), pages 3427-3432. IEEE. DOI: 10.1109/icpr.2018.8546027.
Roweis, S. T. and Saul, L. K. (2000). Nonlinear dimensionality reduction by locally linear embedding. science, 290(5500):2323-2326. DOI: 10.1126/science.290.5500.2323.
Shahroudy, A., Liu, J., Ng, T.-T., and Wang, G. (2016). Ntu rgb+ d: A large scale dataset for 3d human activity analysis. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1010-1019. DOI: 10.1109/cvpr.2016.115.
Souza, L. S., Sogi, N., Gatto, B. B., Kobayashi, T., and Fukui, K. (2023). Grassmannian learning mutual subspace method for image set recognition. Neurocomputing, 517:20-33. DOI: 10.1016/j.neucom.2022.10.040.
Sun, Y., Chen, Y., Wang, X., and Tang, X. (2014). Deep learning face representation by joint identification-verification. Advances in neural information processing systems, 27. DOI: 10.48550/arxiv.1406.4773.
Sun, Z., Hu, Z.-P., Chiong, R., Wang, M., and He, W. (2018). Combining the kernel collaboration representation and deep subspace learning for facial expression recognition. Journal of circuits, systems and computers, 27(08):1850121. DOI: 10.1142/s0218126618501219.
Tenenbaum, J. B., Silva, V. d., and Langford, J. C. (2000). A global geometric framework for nonlinear dimensionality reduction. science, 290(5500):2319-2323. DOI: 10.1126/science.290.5500.2319.
Tian, L., Wang, Z., He, B., He, C., Wang, D., and Li, D. (2021). Knowledge distillation of grassmann manifold network for remote sensing scene classification. Remote Sensing, 13(22):4537. DOI: 10.3390/rs13224537.
Turan, C., Zhao, R., Lam, K.-M., and He, X. (2021). Subspace learning for facial expression recognition: an overview and a new perspective. APSIPA Transactions on Signal and Information Processing, 10:e1. DOI: 10.1017/atsip.2020.27.
Vemulapalli, R., Arrate, F., and Chellappa, R. (2014). Human action recognition by representing 3d skeletons as points in a lie group. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 588-595. DOI: 10.1109/cvpr.2014.82.
Vemulapalli, R. and Chellapa, R. (2016). Rolling rotations for recognizing human actions from 3d skeletal data. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4471-4479. DOI: 10.1109/cvpr.2016.484.
Wang, Q., Gao, J., and Li, H. (2017). Grassmannian manifold optimization assisted sparse spectral clustering. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5258-5266. DOI: 10.1109/cvpr.2017.335.
Wang, R., Wu, X.-J., Chen, Z., Xu, T., and Kittler, J. (2022). Learning a discriminative spd manifold neural network for image set classification. Neural Networks, 151:94-110. DOI: 10.1016/j.neunet.2022.03.012.
Wang, R., Wu, X.-J., Chen, Z., Xu, T., and Kittler, J. (2023a). Dreamnet: A deep riemannian manifold network for spd matrix learning. In Computer Vision – ACCV 2022: 16th Asian Conference on Computer Vision, Macao, China, December 4–8, 2022, Proceedings, Part VI, page 646–663, Berlin, Heidelberg. Springer-Verlag. DOI: 10.1007/978-3-031-26351-4_39.
Wang, R., Wu, X.-J., Xu, T., Hu, C., and Kittler, J. (2023b). U-spdnet: An spd manifold learning-based neural network for visual classification. Neural networks, 161:382-396. DOI: 10.1016/j.neunet.2022.11.030.
Xu, C., Govindarajan, L. N., Zhang, Y., and Cheng, L. (2017). Lie-x: Depth image based articulated object pose estimation, tracking, and action recognition on lie groups. International Journal of Computer Vision, 123:454-478. DOI: 10.1007/s11263-017-0998-6.
Yang, M. and Li, F. (2017). Lie group impression for deep learning. In 2017 International Smart Cities Conference (ISC2), pages 1-5. DOI: 10.1109/ISC2.2017.8090853.
Zhiwu and Van Gool, L. (2017). A riemannian network for spd matrix learning. In Proceedings of the AAAI conference on artificial intelligence, volume 31. DOI: 10.1609/aaai.v31i1.10866.
Zhiwu, Wan, C., Probst, T., and Van Gool, L. (2017). Deep learning on lie groups for skeleton-based action recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 6099-6108. DOI: 10.1109/cvpr.2017.137.
Zhiwu, Wu, J., and Van Gool, L. (2018). Building deep networks on grassmann manifolds. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 32. DOI: 10.1609/aaai.v32i1.11725.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Stéfane Rêgo Gandra, Bernardo Bentes Gatto, Eulanda Miranda dos Santos
This work is licensed under a Creative Commons Attribution 4.0 International License.
