Pith. sign in

REVIEW 3 major objections 6 minor 41 references

A Convolutional Neural Network with Mapping Layers for Hyperspectral Image Classification

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read For hyperspectral image classification, fixed mapping layers from the Tucker decomposition of the average training patch let a shallow 3-D CNN beat deeper baselines in accuracy while training much faster.

desk verdict The mapping-layer idea is genuinely new and the ablation is solid, but the headline accuracy margin over SSRN rests on an overlapping-patch split that leaks spatial context, so the supremacy claim is not yet established. read the letter →

arxiv 1908.09526 v1 pith:QYOWCZBG submitted 2019-08-26 eess.IV cs.CV

classification eess.IVcs.CV
keywords hyperspectralimageclassificationconvolutionalneuralnetworktensordecompositionTuckermappinglayerdimensionreductionspectral-spatialfeatureextraction3-Dconvolution
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to show that hyperspectral image classification can be made both more accurate and much faster by inserting fixed "mapping layers" ahead of a small convolutional network. These layers are derived from a Tucker decomposition of the average training patch and project each input cube into a low-dimensional subspace while keeping its spectral-spatial structure. With only two convolutional layers, the resulting MCNN reports overall accuracies of 98.3% on Indian Pines, 99.5% on University of Pavia, and 99.3% on Salinas, exceeding deeper baselines and cutting training time sharply. The reason this matters is that deep models for hyperspectral data tend to lose accuracy as layers are added and are slow to train, so a shallow network with a principled preprocessing step could be a practical alternative.

What carries the argument

The load-bearing object is the mapping layer: each layer multiplies the input tensor along one mode by a factor matrix $U^{(n)}\in\mathbb{R}^{I_n\times R_n}$ obtained by solving the Tucker decomposition problem for the average training patch with alternating least squares and SVD. Three such layers, one per mode, reduce the spatial and spectral dimensions to $R_1$, $R_2$, $R_3$ and output a small energy-concentrated tensor. Because these kernels are fixed and not updated by back-propagation, the mapping section costs almost no training time, and the small 7x7x40 cube lets two 3-D convolutional layers do the feature extraction that deeper networks need many layers to achieve.

What would settle it

Take a hyperspectral dataset whose classes occupy noticeably different spectral subspaces, build the mapping kernels once from the global average patch and once from per-class average patches, and compare per-class accuracies; if the per-class kernels clearly outperform the global one, the single-average assumption is what limits the method.

Watch

Extended reading notes

Core claim

The central claim is that a neural network for hyperspectral image classification does not need many trainable layers if the input patch is first projected by fixed multilinear mapping kernels. The kernels are the three factor matrices from a Tucker decomposition of the single tensor obtained by averaging all training patches, and they reduce each 13x13x200 patch to a 7x7x40 cube, concentrating most of the energy while preserving the cube structure. A small 3-D convolutional section then extracts spectral-spatial features, and two fully connected layers classify them. The paper reports that this MCNN achieves the highest overall accuracy among the compared methods on all three benchmark datasets and that its total training time is substantially lower than that of the deep residual and deformable baselines.

Load-bearing premise

The whole method rests on the assumption that one tensor formed by averaging all training patches captures the common low-dimensional subspace of every patch well enough that the fixed projection keeps class-discriminative information while removing redundancy.

Editorial extensions

If this is right

  • Classification accuracy on the three tested datasets rises to 98.3%, 99.5%, and 99.3%, all above the compared deep baselines.
  • Total training time drops to tens of seconds on Indian Pines and Salinas, roughly an order of magnitude faster than the residual and deformable baselines.
  • Replacing the mapping layers with PCA or per-patch tensor decomposition lowers accuracy and increases preprocessing time, indicating the averaged-patch projection is the source of the gain.
  • A network with only two convolutional layers avoids the accuracy degradation that the paper associates with deeper hyperspectral CNNs.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same fixed mapping-layer design could be transferred to other tensor-structured inputs, such as video cubes or medical volume data, where a single average tensor may define a shared subspace for all samples.
  • A natural stress test is to replace the single global average patch with class-wise average patches; if per-class kernels improve accuracy on datasets with spectrally distinct classes, the global averaging assumption is the limiting factor.
  • Because the mapping kernels are computed once, one could update them incrementally as new labeled samples arrive, turning the architecture into a cheap few-shot or active-learning classifier.
  • Combining the mapping layers with residual connections may address the gradient-vanishing limitation the paper itself notes, potentially allowing deeper convolutional sections without losing accuracy.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The manuscript proposes MCNN, a hyperspectral image classifier in which three fixed 'mapping layers' are constructed by a higher-order Tucker decomposition (ALS/HOOI) of the average training patch, reducing each 13×13×B input to 7×7×40 before two 3-D convolutional layers and two fully connected layers. The mapping weights are non-trainable, which the authors argue saves training time and avoids accuracy degradation with depth. On Indian Pines, University of Pavia, and Salinas, the method is reported to achieve the highest overall accuracy among SVM, EPF, two 3DCNNs, DHCNet, and SSRN (98.3%, 99.5%, 99.3%) with large reductions in total training time. The paper includes an ablation on Indian Pines comparing the mapping layers with PCA, raw data, and per-patch Tucker decomposition, plus runtime tables.

Significance. If the headline accuracy claim survives a leakage-free evaluation, the paper's contribution is a simple and inexpensive preprocessing module: fixed multilinear projections from a single averaged training patch, combined with a compact 3-D CNN, would be a practically useful baseline for HSI classification. The paper has strengths: three standard datasets; comparisons against six methods including two recent deep networks; multiple-run statistics with standard deviations; an ablation isolating the mapping-layer component; and a time comparison. The main weakness is that the experimental protocol does not exclude spatial overlap between training and test patches, so the central accuracy claim is not yet established.

major comments (3)
  1. [IV-A (data split) and III-A/Algorithm 1] The random 20/10/70 split of 13×13 patches is performed on patch indices, not on disjoint pixels. Neighboring patches overlap by up to 12 of 13 pixels in each spatial direction, so a large fraction of test patch content is present in the training set. Because the mapping kernels are computed from the average of all training patches (Algorithm 1, line 2), the unsupervised projection itself can absorb spatial/spectral information that overlaps the test set. This inflates the reported OA for all patch-based methods and may preferentially inflate MCNN. The margins over SSRN are small (IP: 98.3 vs 97.4; UP: 99.5 vs 99.3), so the claimed supremacy is not established. Please report results on spatially disjoint train/test splits (e.g., disjoint image blocks or a buffer zone between train and test regions), and quantify the overlap of the current split or justify why it does not bias the comparison.
  2. [IV-D and IV-C] The PCA protocol is ambiguous and, as written, unfair. In Section IV-D the authors state that 'PCA is employed as a preprocessing method on training dataset and testing dataset separately'; fitting PCA on the test set separately leaks test-set statistics into the features. If this procedure was also used for the 3DCNN1/3DCNN2/SSRN baselines in Tables V–VII, their accuracies are not comparable to MCNN. The comparison must fit PCA on the training split only and apply the same transformation to validation/test splits; the exact number of components, data centering, and whether PCA is applied per-patch or to the full image should be stated.
  3. [III-B and Fig. 2] The architecture description is internally inconsistent. For Indian Pines, the input to the first convolutional layer is 7×7×40 and the kernel is 5×5×10 with stride (1,1,5). Without padding this produces a 3×3×7×64 output, not the stated 5×5×54×64; padding, dilation, and channel layout are not specified. Since the comparison with 3DCNN1 is justified by 'the same layers and architectures except mapping layers', this discrepancy must be resolved with an exact layer-by-layer configuration.
minor comments (6)
  1. [IV-C] The claim that MCNN 'increases the OA compared with 3DCNN1 method by about 2%−7%' is inconsistent with Tables V–VII: the differences are 9.4, 2.0, and 4.4 percentage points on IP, UP, and Salinas, respectively.
  2. [Algorithm 1 and Section III-A] Use a distinct symbol (e.g., \bar{X}) for the averaged training patch; currently X denotes both the training patch tensor and its average, which is confusing.
  3. [Equation (4)] Equation (4) contains an index error: after the mode-n product the mode-n index should be the column index of U, and the summation should run over the original mode-n index. Please correct.
  4. [End of Section IV-D] The sentence 'The disadvantage of our MCNN is that it only avoids the gradient vanishing problem, this problem can only be sovled by residual network so far' is unclear and contains a typo; please state precisely what limitation is intended (e.g., the fixed mapping layers are not learned and may not adapt to heterogeneous patches).
  5. [Section III-A and Section V] The claim that 'most energy of the input is preserved' is never quantified. Report the retained Tucker energy (e.g., sum of retained core entries divided by total energy) for the chosen (R1,R2,R3) on each dataset, or soften the claim.
  6. [Tables VIII-X] The DHCNet rows list no per-epoch time or epoch count, so the total training time cannot be interpreted; specify the protocol used to obtain these numbers.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the mapping layers are unsupervised projections and the reported accuracies are measured on held-out pixels.

full rationale

The paper's derivation chain is self-contained and empirical: the mapping kernels U^(1), U^(2), U^(3) are obtained by a Tucker decomposition of the averaged training patch (Algorithm 1, Section III-A), using only unlabeled tensor structure, not class labels. The classification accuracies reported in Tables V–VII are measured on a held-out test split after training the CNN with backpropagation, so no test quantity is used to construct the mapping layers, and no fitted parameter is renamed as a prediction. The claim that the mapping layers preserve most of the input energy follows from the well-known optimality of singular vectors for low-rank reconstruction; the paper does not quantify it, which is an unsubstantiated assertion rather than a circular reduction. Comparisons against SSRN, DHCNet, 3DCNN, EPF, and SVM are standard external benchmark comparisons, and the paper's self-citations are ordinary references to prior methods, not load-bearing uniqueness arguments. The potential concern that randomly split overlapping patches may cause spatial leakage is an experimental-validity issue, not a circularity of the argument, because the reported numbers are measurements rather than consequences of the mapping-layer construction by definition. No equation in the paper reduces to its own inputs, and no specific circular step could be exhibited.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central result depends on a handful of validation-selected hyperparameters and the domain assumption that an averaged training patch adequately represents the common subspace. No new physical entities are introduced.

free parameters (5)
  • R1, R2, R3 (mapping layer output sizes) = (7,7,40) for Indian Pines and Salinas; (7,7,20) for University of Pavia
    Selected by validation on each dataset (Tables II-IV); the central accuracy numbers depend on these choices, and the paper recommends different values for different datasets.
  • Learning rate = 0.001 for Indian Pines; 0.003 for University of Pavia and Salinas
    Chosen by grid search over (0.01, 0.003, 0.001, 0.0003, 0.0001) using validation accuracy (Fig. 7).
  • Batch size = 30 for all datasets
    Grid searched over (20, 30, 40), validation-selected.
  • Training epochs = 30 for all datasets
    Early stopping via validation accuracy (Fig. 8).
  • ALS stopping threshold = 0.01 on difference of core tensors
    Algorithm 1, line 5; a hand-chosen tolerance for mapping-layer convergence.
assumptions (4)
  • domain assumption Tucker decomposition of the average training patch identifies a subspace that is representative of all training patches.
    Section III-A: 'In this way, our mapping layer can reduce the common redundancy of all training patches'. No evidence that a single averaged patch captures the common subspace.
  • domain assumption The energy concentration property of truncated singular vectors guarantees that the mapping keeps the most discriminative information for classification.
    Section III-A and IV state 'most energy of the input patch is still kept', but no energy fraction is quantified and no link to class discrimination is proven.
  • domain assumption Random pixel-level patch splitting gives a valid estimate of generalization.
    Section IV-A: 20% patches train, 10% validation, 70% test. Because patches overlap by up to 12 pixels in a 13x13 window, train and test patches share pixels, likely inflating accuracy.
  • standard math Standard multilinear algebra operations (mode-n product, SVD) behave as described.
    Used throughout Section II; no issues.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Convolutional Neural Network with Mapping Layers for Hyperspectral Image Classification." pith.science (2026). https://pith.science/paper/QYOWCZBG

@misc{pith2026190809526,
  author       = {Pith},
  title        = {Pith review of: A Convolutional Neural Network with Mapping Layers for Hyperspectral Image Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QYOWCZBG}},
  note         = {Machine review of arXiv:1908.09526}
}
read the original abstract

In this paper, we propose a convolutional neural network with mapping layers (MCNN) for hyperspectral image (HSI) classification. The proposed mapping layers map the input patch into a low dimensional subspace by multilinear algebra. We use our mapping layers to reduce the spectral and spatial redundancy and maintain most energy of the input. The feature extracted by our mapping layers can also reduce the number of following convolutional layers for feature extraction. Our MCNN architecture avoids the declining accuracy with increasing layers phenomenon of deep learning models for HSI classification and also saves the training time for its effective mapping layers. Furthermore, we impose the 3-D convolutional kernel on convolutional layer to extract the spectral-spatial features for HSI. We tested our MCNN on three datasets of Indian Pines, University of Pavia and Salinas, and we achieved the classification accuracy of 98.3%, 99.5% and 99.3%, respectively. Experimental results demonstrate that the proposed MCNN can significantly improve the classification accuracy and save much time consumption.

Figures

Figures reproduced from arXiv: 1908.09526 by the authors.

Figure 1
Figure 1. The mode-1 matricization is obtained by flattening the tensor along [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The architecture of our proposed MCNN [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. The comparison of two kinds of convolutional kernels. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: (a) Three-band color composite of the Indian Pines image; (b) ground [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: (a) Three-band color composite of the UP image; (b) ground truth of [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 7
Figure 7. Figure 7: The validation results of 15 models on three datasets, each model is the [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: The validation accuracy varies with the epoch number on the best [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: Classification results of all methods for Indian Pines, (a) ground truth; [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 11
Figure 11. Figure 11: Classification results of all methods for Salinas, (a) ground truth; (b)- [PITH_FULL_IMAGE:figures/full_fig_p011_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 37 canonical work pages

  1. [1]

    Modern trends in hyperspectral image analysis: A review,

    M. J. Khan, H. S. Khan, A. Yousaf, K. Khurshid, and A. Abbas, “Modern trends in hyperspectral image analysis: A review,” IEEE Access, vol. 6, pp. 14 118–14 129, Mar. 2018

  2. [2]

    Semi-supervised learning through label propagation on geodesics,

    M. Fan, X. Zhang, L. Du, L. Chen, and D. Tao, “Semi-supervised learning through label propagation on geodesics,” IEEE Trans. Cybern., vol. 48, no. 5, pp. 1486–1499, Jul. 2018

  3. [3]

    Extended random walker-based classification of hyperspectral images,

    X. Kang, S. Li, L. Fang, M. Li, and J. A. Benediktsson, “Extended random walker-based classification of hyperspectral images,” IEEE Trans. Geosci. Remote Sens. , vol. 53, no. 1, pp. 144–153, Dec. 2015

  4. [4]

    Salient band selection for hyperspectral image classification via manifold ranking,

    Q. Wang, J. Lin, and Y . Yuan, “Salient band selection for hyperspectral image classification via manifold ranking,” IEEE Trans. Neural Netw. Learn. Syst., vol. 27, no. 6, pp. 1279–1289, Jun. 2016

  5. [5]

    Similarity-based unsupervised band selection for hyperspectral image analysis

    Q. Du and H. Yang, “Similarity-based unsupervised band selection for hyperspectral image analysis.” IEEE Geosci. Remote Sens. Lett. , vol. 5, no. 4, pp. 564–568, Oct. 2008

  6. [6]

    Hyperspectral image classification via multitask joint sparse representation and stepwise mrf optimization,

    Y . Yuan, J. Lin, and Q. Wang, “Hyperspectral image classification via multitask joint sparse representation and stepwise mrf optimization,” IEEE Trans. Cybern. , vol. 46, no. 12, pp. 2966–2977, Dec. 2016

  7. [7]

    Superpca: A superpixelwise pca approach for unsupervised feature extraction of hyperspectral imagery,

    J. Jiang, J. Ma, C. Chen, Z. Wang, Z. Cai, and L. Wang, “Superpca: A superpixelwise pca approach for unsupervised feature extraction of hyperspectral imagery,” IEEE Trans. Geosci. Remote Sens. , vol. 56, no. 8, pp. 4581–4593, Aug. 2018

  8. [8]

    Limitations of principal components analysis for hyperspectral target recognition,

    S. Prasad and L. M. Bruce, “Limitations of principal components analysis for hyperspectral target recognition,” IEEE Geosci. Remote Sens. Lett., vol. 5, no. 4, pp. 625–629, Oct. 2008

Show all 41 references
  1. [9]

    Unsupervised feature ex- traction based on a mutual information measure for hyperspectral image classification,

    M. A. Hossain, M. Pickering, and X. Jia, “Unsupervised feature ex- traction based on a mutual information measure for hyperspectral image classification,” in Proc. IEEE Int. Geosci. Remote Sens. Symp. (IGARSS), 2011, pp. 1720–1723

  2. [10]

    Dimensionality reduction via regression in hyperspectral imagery,

    V . Laparra, J. Malo, and G. Camps-Valls, “Dimensionality reduction via regression in hyperspectral imagery,” IEEE J. Sel. Topics Signal Process., vol. 9, no. 6, pp. 1026–1036, Sep. 2015

  3. [11]

    Classification of hyperspectral images with regularized linear discriminant analysis,

    T. V . Bandos, L. Bruzzone, and G. Camps-Valls, “Classification of hyperspectral images with regularized linear discriminant analysis,” IEEE Trans. Geosci. Remote Sens. , vol. 47, no. 3, pp. 862–873, Mar. 2009

  4. [12]

    A modified locality-preserving projection approach for hyperspectral image classification,

    Y . Zhai, L. Zhang, N. Wang, Y . Guo, Y . Cen, T. Wu, and Q. Tong, “A modified locality-preserving projection approach for hyperspectral image classification,” IEEE Geosci. Remote Sens. Lett. , vol. 13, no. 8, pp. 1059–1063, Aug. 2016

  5. [13]

    Discover latent discriminant information for dimensional- ity reduction: Non-negative sparseness preserving embedding,

    W. K. Wong, “Discover latent discriminant information for dimensional- ity reduction: Non-negative sparseness preserving embedding,” Pattern recognition, vol. 45, no. 4, pp. 1511–1523, Apr. 2012

  6. [14]

    Locality and structure regularized low rank representation for hyperspectral image classification,

    Q. Wang, X. He, and X. Li, “Locality and structure regularized low rank representation for hyperspectral image classification,” IEEE Trans. Geosci. Remote Sens. , vol. 57, no. 2, pp. 911–923, Feb. 2018

  7. [15]

    Spectral–spatial hyperspectral image classification with edge-preserving filtering,

    X. Kang, S. Li, and J. A. Benediktsson, “Spectral–spatial hyperspectral image classification with edge-preserving filtering,” IEEE Trans. Geosci. Remote Sens., vol. 52, no. 5, pp. 2666–2677, May 2014

  8. [16]

    Kernel-based methods for hyperspec- tral image classification,

    G. Camps-Valls and L. Bruzzone, “Kernel-based methods for hyperspec- tral image classification,” IEEE Trans. Geosci. Remote Sens. , vol. 43, no. 6, pp. 1351–1362, Jun. 2005

  9. [17]

    Classification and feature extraction for remote sensing images from urban areas based on morphological transformations,

    J. A. Benediktsson, M. Pesaresi, and K. Amason, “Classification and feature extraction for remote sensing images from urban areas based on morphological transformations,” IEEE Trans. Geosci. Remote Sens. , vol. 41, no. 9, pp. 1940–1949, Sep. 2003

  10. [18]

    Composite kernels for hyperspectral image classification,

    G. Camps-Valls, L. Gomez-Chova, J. Mu ˜noz-Mar´ı, J. Vila-Franc ´es, and J. Calpe-Maravilla, “Composite kernels for hyperspectral image classification,” IEEE Geosci. Remote Sens. Lett., vol. 3, no. 1, pp. 93–97, Jan. 2006

  11. [19]

    Video segmentation with superpixels,

    F. Galasso, R. Cipolla, and B. Schiele, “Video segmentation with superpixels,” in Proc. ACCV. Springer, 2012, pp. 760–774

  12. [20]

    Representation learning: A review and new perspectives,

    Y . Bengio, A. Courville, and P. Vincent, “Representation learning: A review and new perspectives,” IEEE Trans. Pattern Anal. Mach. Intell. , vol. 35, no. 8, pp. 1798–1828, Aug. 2013

  13. [21]

    Deep hierarchies in the primate visual cortex: What can we learn for computer vision?

    N. Kruger, P. Janssen, S. Kalkan, M. Lappe, A. Leonardis, J. Piater, A. J. Rodriguez-Sanchez, and L. Wiskott, “Deep hierarchies in the primate visual cortex: What can we learn for computer vision?” IEEE Trans. Pattern Anal. Mach. Intell. , vol. 35, no. 8, pp. 1847–1871, Aug. 2013

  14. [22]

    Deep learning-based classification of hyperspectral data,

    Y . Chen, Z. Lin, X. Zhao, G. Wang, and Y . Gu, “Deep learning-based classification of hyperspectral data,” IEEE J. Sel. Topics Appl. Earth Observ. Remote Sens. , vol. 7, no. 6, pp. 2094–2107, Jun. 2014

  15. [23]

    Spectral–spatial feature extraction for hsi classification based on supervised hypergraph and sample expanded cnn,

    Y . Kong, X. Wang, and Y . Cheng, “Spectral–spatial feature extraction for hsi classification based on supervised hypergraph and sample expanded cnn,” IEEE J. Sel. Topics Appl. Earth Observ. Remote Sens. , vol. 11, no. 11, pp. 4128–4140, Nov. 2018

  16. [24]

    Cnn-based multilayer spatial–spectral feature fusion and sample aug- mentation with local and nonlocal constraints for hyperspectral image classification,

    J. Feng, J. Chen, L. Liu, X. Cao, X. Zhang, L. Jiao, and T. Yu, “Cnn-based multilayer spatial–spectral feature fusion and sample aug- mentation with local and nonlocal constraints for hyperspectral image classification,” IEEE J. Sel. Topics Appl. Earth Observ. Remote Sens. , vo...

  17. [25]

    Multisource remote sensing data classification based on convolutional neural network,

    X. Xu, W. Li, Q. Ran, Q. Du, L. Gao, and B. Zhang, “Multisource remote sensing data classification based on convolutional neural network,” IEEE Trans. Geosci. Remote Sens. , vol. 56, no. 2, pp. 937–949, Feb. 2017

  18. [26]

    Scene classification with recurrent attention of vhr remote sensing images,

    Q. Wang, S. Liu, J. Chanussot, and X. Li, “Scene classification with recurrent attention of vhr remote sensing images,” IEEE Trans. Geosci. Remote Sens., vol. 57, no. 2, pp. 1156–1167, Feb. 2019

  19. [27]

    Deep feature extraction and classification of hyperspectral images based on convolutional neural networks,

    Y . Chen, H. Jiang, C. Li, X. Jia, and P. Ghamisi, “Deep feature extraction and classification of hyperspectral images based on convolutional neural networks,” IEEE Trans. Geosci. Remote Sens., vol. 54, no. 10, pp. 6232– 6251, Oct. 2016

  20. [28]

    Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion,

    P. Vincent, H. Larochelle, I. Lajoie, Y . Bengio, and P.-A. Manzagol, “Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion,” J. Mach. Learn. Res. , vol. 11, no. Dec, pp. 3371–3408, Dec. 2010

  21. [29]

    Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations,

    H. Lee, R. Grosse, R. Ranganath, and A. Y . Ng, “Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations,” in Proc. Int’l Conf. Mach. Learn. ACM, 2009, pp. 609–616

  22. [30]

    Going deeper with contextual cnn for hyperspec- tral image classification,

    H. Lee and H. Kwon, “Going deeper with contextual cnn for hyperspec- tral image classification,” IEEE Trans. Image Process. , vol. 26, no. 10, pp. 4843–4855, Oct. 2017

  23. [31]

    Spectral-spatial residual network for hyperspectral image classification: A 3-d deep learning framework,

    Z. Zhong, J. Li, Z. Luo, and M. Chapman, “Spectral-spatial residual network for hyperspectral image classification: A 3-d deep learning framework,” IEEE Trans. Geosci. Remote Sens., vol. 56, no. 2, pp. 847– 858, Feb. 2018

  24. [32]

    High-order possibilistic c- means algorithms based on tensor decompositions for big data in iot,

    Q. Zhang, L. T. Yang, Z. Chen, and P. Li, “High-order possibilistic c- means algorithms based on tensor decompositions for big data in iot,” Inf. Fusion, vol. 39, pp. 72–80, Jan. 2018

  25. [33]

    Foundations of the parafac procedure: Models and conditions for an

    R. A. Harshman, “Foundations of the parafac procedure: Models and conditions for an” explanatory” multimodal factor analysis,” UCLA Working Papers in Phonetics, Dec. 1970

  26. [34]

    A multilinear singular value decomposition,

    L. De Lathauwer, B. De Moor, and J. Vandewalle, “A multilinear singular value decomposition,” SIAM J. Matrix Anal. Appl. , vol. 21, no. 4, pp. 1253–1278, Apr. 2000

  27. [35]

    Some mathematical notes on three-mode factor analysis,

    L. R. Tucker, “Some mathematical notes on three-mode factor analysis,” Psychometrika, vol. 31, no. 3, pp. 279–311, 1966

  28. [36]

    Very deep convolutional networks for large-scale image recognition,

    K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv:1409.1556 , 2014. UNDER REVIEW ON IEEE TRANS. GEOSCI. REMOTE SEN. 13

  29. [37]

    Deep learning,

    Y . LeCun, Y . Bengio, and G. Hinton, “Deep learning,”Nature, vol. 521, no. 7553, p. 436, Sep. 2015

  30. [38]

    Human-level control through deep reinforcement learning,

    V . Mnih, K. Kavukcuoglu, D. Silver, A. A. Rusu, J. Veness, M. G. Bellemare, A. Graves, M. Riedmiller, A. K. Fidjeland, G. Ostrovski et al. , “Human-level control through deep reinforcement learning,” Nature, vol. 518, no. 7540, pp. 529–533, Feb. 2015

  31. [39]

    Deformable convolutional neural networks for hyperspectral image classification,

    J. Zhu, L. Fang, and P. Ghamisi, “Deformable convolutional neural networks for hyperspectral image classification,” IEEE Geosci. Remote Sens. Lett., vol. 15, no. 8, pp. 1254–1258, Aug. 2018

  32. [40]

    Spectral–spatial classification of hyper- spectral imagery with 3d convolutional neural network,

    Y . Li, H. Zhang, and Q. Shen, “Spectral–spatial classification of hyper- spectral imagery with 3d convolutional neural network,” Remote Sens., vol. 9, no. 1, p. 67, 2017

  33. [41]

    Libsvm: a library for support vector machines,

    C.-C. Chang and C.-J. Lin, “Libsvm: a library for support vector machines,” ACM TIST, vol. 2, no. 3, p. 27, Apr. 2011. Rui Li received the B. S. and M. S. degree from Xidian University, Xi’an, P. R. China, in 2012 and 2015 respectively. He is currently working towards the Ph. ...

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.