Abstract
Few-shots learning is a popular transfer learning paradigm and leverages an additional source of side information to compensate for the limited labelled training exemplars in application domains like healthcare, military and space. Most few-shots learners do not capture useful data manifolds and have numerous trainable parameters. This large parameter space makes the inference computationally expensive and energy-draining. Consequently, these models are not robust and cannot be deployed on devices with limited computation (energy) faculties. This work proposes NucNormFSL, a novel nuclear norm-induced lightweight relation network, for the few-shots classification of hyperspectral images. The embedding and relation modules in the proposed network are trained end-to-end by minimizing a dictionary learning-based loss function with only a few trainable parameters. Additionally, the embedding module loss function is regularized using a nuclear norm to give low-ranked solutions that are robust to environmental noise; lastly, a relative reconstruction loss metric is introduced to quantify the embedding’s robustness to noise. Experiments are conducted on four benchmark hyperspectral datasets, namely, Indian Pines, Pavia Center, Pavia University and Salinas dataset; the relative reconstruction loss values computed confirm the robustness of the embeddings and hence the proposed network to environmental noise. Additionally, the proposed network’s performance is compared with the baseline (without a nuclear norm term in its embedding loss function) model. The proposed approach beats the baseline for most few-shots settings and datasets and remains competitive with the state-of-the-art despite being severely lightweight. In this way, the proposed network is futuristic, lightweight and immune to noise; consequently, it can be deployed in noisy environments on devices with limited computation facilities.
Similar content being viewed by others
Data Availability
The experimental data will be made available from the corresponding author upon reasonable request
Code Availability
The codes will be made publicly available upon acceptance.
References
Alajaji D, Alhichri HS, Ammour N, Alajlan N (2020) Few-shot learning for remote sensing scene classification. In: 2020 Mediterranean and Middle-East Geoscience and Remote Sensing Symposium (M2GARSS), pp. 81–84. IEEE
Alloghani M, Al-Jumeily D, Mustafina J, Hussain A, Aljaaf AJ (2020) A systematic review on supervised and unsupervised machine learning algorithms for data science. Supervised and unsupervised learning for data science, pp 3–21
Alomari A, Idris N, Sabri AQM, Alsmadi I (2022) Deep reinforcement and transfer learning for abstractive text summarization: A review. Comput Speech Language 71:101276
Bai J, Huang S, Xiao Z, Li X, Zhu Y, Regan AC, Jiao L (2022) Few-shot hyperspectral image classification based on adaptive subspaces and feature transformation. IEEE Trans Geosci Remote Sens 60:1–17
Bau TC, Sarkar S, Healey G (2010) Hyperspectral region classification using a three-dimensional gabor filterbank. IEEE Trans Geosci Remote Sens 48 (9):3457–3464. https://doi.org/10.1109/TGRS.2010.2046494
Bhangale KB, Mohanaprasad K (2021) A review on speech processing using machine learning paradigm. Int J Speech Technol 24(2):367–388
Bing L, Xibing Z, Xiong T, Anzhu Y, Wenyue G (2020) A deep few-shot learning algorithm for hyperspectral image classification. Acta Geodaetica et Cartographica Sinica 49(10):1331
Chen Y, Lin Z, Zhao X, Wang G, Gu Y (2014) Deep learning-based classification of hyperspectral data. Sel Top Appl Earth Obs Remote Sens, IEEE J 7:2094–2107. https://doi.org/10.1109/JSTARS.2014.2329330
Chen Y, Zhao X, Jia X (2015) Spectral–spatial classification of hyperspectral data based on deep belief network. IEEE J Sel Top Appl Earth Obs Remote Sens 8(6):2381–2392. https://doi.org/10.1109/JSTARS.2015.2388577
Cheng G, Cai L, Lang C, Yao X, Chen J, Guo L, Han j. (2021) Spnet: Siamese-prototype network for few-shot remote sensing image scene classification. IEEE Trans Geosci Remote Sens 60:1–11
Cremer CZ (2021) Deep limitations? examining expert disagreement over deep learning. Progress Artif Intell 10(4):449–464
Deng B, Jia S, Shi D (2020) Deep metric learning-based feature embedding for hyperspectral image classification. IEEE Trans Geosci Remote Sens 58 (2):1422–1435. https://doi.org/10.1109/TGRS.2019.2946318
Dong S, Wang P, Abbas K (2021) A survey on deep learning and its applications. Comput Sci Rev 40:100379
Falco N, Bruzzone L, Benediktsson JA (2014) An ica based approach to hyperspectral image feature reduction. In: 2014 IEEE Geoscience and remote sensing symposium, pp 3470–3473. https://doi.org/10.1109/IGARSS.2014.6947229
Fazel M, Hindi H, Boyd SP (2001) A rank minimization heuristic with application to minimum order system approximation. In: Proceedings of the 2001 American control conference.(Cat. No. 01CH37148), vol 6, pp 4734–4739. IEEE
Gao K, Liu B, Yu X, Qin J, Zhang P, Tan X (2020) Deep relation network for hyperspectral image few-shot classification. Remote Sensing 12(6). https://doi.org/10.3390/rs12060923
Gao K, Liu B, Yu X, Zhang P, Tan X, Sun Y (2021) Small sample classification of hyperspectral image using model-agnostic meta-learning algorithm and convolutional neural network. Int J Remote Sens 42(8):3090–3122
Geng C, Huang S-J, Chen S (2020) Recent advances in open set recognition: A survey. IEEE Trans Pattern Anal Mach Intell 43(10):3614–3631
Gong Z, Zhong P, Yu Y, Hu W, Li S (2019) A cnn with multiscale convolution and diversified metric for hyperspectral image classification. IEEE Trans Geosci Remote Sens 57(6):3599–3618. https://doi.org/10.1109/TGRS.2018.2886022
Harikiran J, Reddy TS (2022) An outlook: machine learning in hyperspectral image classification and dimensionality reduction techniques. Journal of Spectral Imaging 11
He L, Chen X (2016) A three-dimensional filtering method for spectral-spatial hyperspectral image classification. In: 2016 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), pp 2746–2748. https://doi.org/10.1109/IGARSS.2016.7729709
Hu Y, Huang Y, Wei G, Zhu K (2022) Heterogeneous few-shot learning with knowledge distillation for hyperspectral image classification. In: 2022 2nd International Conference on Consumer Electronics and Computer Engineering (ICCECE), pp 601–604. IEEE
Hu L, Luo X, Wei Y (2020) Hyperspectral image classification of convolutional neural network combined with valuable samples. J Phys Conf Ser 1549 (5):052011. https://doi.org/10.1088/1742-6596/1549/5/052011
Hu Z, Nie F, Tian L, Wang R, Li X (2018) A comprehensive survey for low rank regularization. arXiv:1808.04521
Huang W, Yuan Z, Yang A, Tang C, Luo X (2021) Tae-net: task-adaptive embedding network for few-shot remote sensing scene classification. Remote Sensing 14(1):111
Iwata T, Tanaka Y (2022) Few-shot learning for spatial regression via neural embedding-based gaussian processes. Machine Learn 111(4):1239–1257
Jia S, Hu J, Zhu J, Jia X, Li Q (2017) Three-dimensional local binary patterns for hyperspectral imagery classification. IEEE Trans Geosci Remote Sens 55(4):2399–2413. https://doi.org/10.1109/TGRS.2016.2642951
Jia S, Jiang S, Lin Z, Li N, Xu M, Yu S (2021) A survey: Deep learning for hyperspectral image classification with few labeled samples. Neurocomputing 448:179–204
Jiao L, Liang M, Chen H, Yang S, Liu H, Cao X (2017) Deep fully convolutional network-based spatial distribution prediction for hyperspectral image classification. IEEE Trans Geosci Remote Sens 55:5585–5599
Le N, Rathour VS, Yamazaki K, Luu K, Savvides M (2021) Deep reinforcement learning in computer vision: a comprehensive survey. Artificial Intelligence Review, pp 1–87
Li X, Cao Z, Zhao L, Jiang J (2021) Alpn: Active-learning-based prototypical network for few-shot hyperspectral imagery classification. IEEE Geosci Remote Sens Lett 19:1–5
Li Z, Liu M, Chen Y, Xu Y, Li W, Du Q (2021) Deep cross-domain few-shot learning for hyperspectral image classification. IEEE Trans Geosci Remote Sens 60:1–18
Liao W, Pizurica A, Philips W, Pi Y (2010) A fast iterative kernel pca feature extraction for hyperspectral images. In: 2010 IEEE International conference on image processing, pp 1317–1320. https://doi.org/10.1109/ICIP.2010.5651670
Licciardi G, Marpu PR, Chanussot J, Benediktsson JA (2012) Linear versus nonlinear pca for the classification of hyperspectral data based on the extended morphological profiles. IEEE Geosci Remote Sens Lett 9(3):447–451. https://doi.org/10.1109/LGRS.2011.2172185
Liu B, Yu X, Yu A, Zhang P, Wan G, Wang R (2019) Deep few-shot learning for hyperspectral image classification. In: IEEE Transactions on geo science and remote sensing, vol 57
Liu Y, Zhang H, Zhang W, Lu G, Tian Q, Ling N (2022) Few-shot image classification: Current status and research trends. Electronics 11 (11):1752
Ma C, Mu X, Zhao P, Yan X (2021) Meta-learning based on parameter transfer for few-shot classification of remote sensing scenes. Remote Sens Lett 12(6):531–541
Mankolli E, Guliashki V (2020) Machine learning and natural language processing: Review of models and optimization problems. In: International Conference on ICT Innovations, pp 71–86. Springer
Mei S, Ji J, Geng Y, Zhang Z, Li X, Du Q (2019) Unsupervised spatial–spectral feature learning by 3d convolutional autoencoder for hyperspectral classification. IEEE Trans Geosci Remote Sens 57:9
Melgani F, Bruzzone L (2004) Classification of hyperspectral remote sensing images with support vector machines. IEEE Trans Geosci Remote Sens 42 (8):1778–1790. https://doi.org/10.1109/TGRS.2004.831865
Mughees A, Tao L (2019) Multiple deep-belief-network-based spectral-spatial classification of hyperspectral images. Tsinghua Sci Technol 24(2):183–194. https://doi.org/10.26599/TST.2018.9010043
Pal D, Bundele V, Banerjee B, Jeppu Y (2021) Spn: Stable prototypical network for few-shot learning-based hyperspectral image classification. IEEE Geosci Remote Sens Lett 19:1–5
Pandey SK, Shekhawat HS, Prasanna SM (2019) Deep learning techniques for speech emotion recognition: A review. In: 2019 29th International Conference Radioelektronika (RADIOELEKTRONIKA), pp 1–6. IEEE
Patel H, Upla KP (2022) A shallow network for hyperspectral image classification using an autoencoder with convolutional neural network. Multimed Tools Appl 81(1):695–714
Quesada-Barriuso P, Argüello F, Heras DB (2014) Spectral–spatial classification of hyperspectral images using wavelets and extended morphological profiles. IEEE J Sel Top Appl Earth Obs Remote Sens 7(4):1177–1185. https://doi.org/10.1109/JSTARS.2014.2308425
Rao M, Tang P, Zhang Z (2019) Spatial–spectral relation network for hyperspectral image classification with limited training samples. IEEE J Sel Top Appl Earth Obs Remote Sens 12(12):5086–5100
Recht B, Fazel M, Parrilo PA (2010) Guaranteed minimum-rank solutions of linear matrix equations via nuclear norm minimization. SIAM review 52 (3):471–501
Ren L, Duan G, Huang T, Kang Z (2022) Multi-local feature relation network for few-shot learning. Neural Comput Applic 34(10):7393–7403
Ren Y, Zhang Y, Wei W, Li L (2014) A spectral-spatial hyperspectral data classification approach using random forest with label constraints. In: 2014 IEEE Workshop on electronics, computer and applications, pp 344–347. https://doi.org/10.1109/IWECA.2014.6845627
Sagar R, Jhaveri R, Borrego C (2020) Applications in security and evasions in machine learning: a survey. Electronics 9(1):97
Sanghvi K, Aralkar A, Sanghvi S, Saha I (2020) A survey on image classification techniques. Available at SSRN 3754116
Sarker IH (2021) Data science and analytics: an overview from data-driven smart computing, decision-making and applications perspective. SN Comput Sci 2(5):1–22
Singh UP, Singh KP, Thakur M (2022) Nucnormzsl: nuclear norm-based domain adaptation in zero-shot learning. Neural Comput Appl 34(3):2353–2374
Tan C, Sun F, Kong T, Zhang W, Yang C, Liu C (2018) A survey on deep transfer learning. In: International conference on artificial neural networks, pp 270–279. Springer
Tang H, Li Y, Han X, Huang Q, Xie W (2019) A spatial–spectral prototypical network for hyperspectral remote sensing image. IEEE Geosci Remote Sens Lett 17(1):167–171
Tong X, Yin J, Han B, Qv H (2020) Few-shot learning with attention-weighted graph convolutional networks for hyperspectral image classification. In: 2020 IEEE International Conference on Image Processing (ICIP), pp 1686–1690. IEEE
Torrey L, Shavlik J (2010) Transfer learning. In: Handbook of research on machine learning applications and trends: algorithms, Methods, and Techniques, pp 242–264. IGI global
Vangara RVB, Vangara SP, Thirupathur V (2020) A survey on natural language processing in context with machine learning. Int J Anal Exp Modal Anal, pp 1390–1395
Wang S, Du B, Zhang D, Wan F (2021) Adversarial prototype learning for hyperspectral image classification IEEE Transactions on Geoscience and Remote Sensing
Wang Y, Liu M, Yang Y, Li Z, Du Q, Chen Y, Li F, Yang H (2021) Heterogeneous few-shot learning for hyperspectral image classification. IEEE Geosci Remote Sens Lett 19:1–5
Wang G, Zheng X, Cheng L, Wan X, Guo Z (2021) Hyperspectral image classification based on improved few shot learning. In: 2021 IEEE International Conference on Electronic Technology, Communication and Information (ICETCI), pp 673–676. IEEE
Yang S, Gao T, Wang J, Deng B, Azghadi MR, Lei T, Linares-Barranco B (2022) Sam: a unified self-adaptive multicompartmental spiking neuron model for learning with working memory. Frontiers in Neuroscience 16
Yang S, Linares-Barranco B, Chen B (2022) Heterogeneous ensemble-based spike-driven few-shot online learning. Frontiers in Neuroscience 16
Yang S, Tan J, Chen B (2022) Robust spike-based continual meta-learning improved by restricted minimum error entropy criterion. Entropy 24(4):455
Yang S, Wang J, Deng B, Azghadi MR, Linares-Barranco B (2021) Neuromorphic context-dependent learning framework with fault-tolerant spike routing. IEEE Trans Neural Netw Learn Syst 33(12):7126–7140
Yu S, Jia S, Xu C (2017) Convolutional neural networks for hyperspectral image classification. Neurocomputing 219:88–98. https://doi.org/10.1016/j.neucom.2016.09.010
Zhang P, Bai Y, Wang D, Bai B, Li Y (2021) Few-shot classification of aerial scene images via meta-learning. Remote Sens 13(1):108
Zhang Y, Li W, Zhang M, Tao R (2022) Dual graph cross-domain few-shot learning for hyperspectral image classification. In: ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 3573–3577. IEEE
Zhang C, Yue J, Qin Q (2020) Deep quadruplet network for hyperspectral image classification with a small number of samples. Remote Sens 12(4):647
Zhang C, Yue J, Qin Q (2020) Global prototypical network for few-shot hyperspectral image classification. IEEE J Sel Top Appl Earth Obs Remote Sens 13:4748–4759
Zhao J, Hu L, Dong Y, Huang L, Weng S, Zhang D (2021) A combination method of stacked autoencoder and 3d deep residual network for hyperspectral image classification. Int J Appl Earth Obs Geoinf 102:102459
Zheng C, Zheng Y (2014) Hyperspectral remote sensing image classification based on combined svm and lda. SPIE Asia Pac Remote Sens
Zhong Z, Li J, Luo Z, Chapman M (2017) Spectral-spatial residual network for hyperspectral image classification: A 3-d deep learning framework. IEEE Trans Geosci Remote Sens 56:847–858. https://doi.org/10.1109/TGRS.2017.2755542
Zhuang F, Qi Z, Duan K, Xi D, Zhu Y, Zhu H, Xiong H, He Q (2020) A comprehensive survey on transfer learning. Proc IEEE 109(1):43–76
Zohuri B, Moghaddam M (2020) Deep learning limitations and flaws. Mod Approaches Mater Sci 2:241–250
Funding
Not Applicable
Author information
Authors and Affiliations
Contributions
All authors have contributed to this research study in different capacities. For instance, the study was conceptualised by Upendra Pratap Singh and Krishna Pratap Singh, while Upendra Pratap Singh performed material preparation, data analysis and experiments. Upendra Pratap Singh wrote the manuscript’s first draft, and Manoj Thakur and Krishna Pratap Singh suggested relevant improvements. The final version of the manuscript was proofread and approved by all the authors.
Corresponding author
Ethics declarations
Ethics approval
Not Applicable
Consent for Participate
Not Applicable
Consent for Publication
Not Applicable
Conflict of Interests
All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Singh, U.P., Singh, K.P. & Thakur, M. A nuclear norm-induced robust and lightweight relation network for few-shots classification of hyperspectral images. Multimed Tools Appl 83, 9279–9306 (2024). https://doi.org/10.1007/s11042-023-15500-z
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-023-15500-z