Hyperspectral classification based on lightweight 3D-CNN with transfer learning.

Standard

Hyperspectral classification based on lightweight 3D-CNN with transfer learning. / Zhang, Haokui ; Li, Ying; Jiang, Yenan ; Wang, Peng ; Shen, Qiang; Shen, Chunhua .

In: IEEE Transactions on Geoscience and Remote Sensing, Vol. 57, No. 8, 01.08.2019, p. 5813-5828.

Research output: Contribution to journalArticle

Harvard

Zhang, H, Li, Y, Jiang, Y, Wang, P, Shen, Q & Shen, C 2019, 'Hyperspectral classification based on lightweight 3D-CNN with transfer learning.' IEEE Transactions on Geoscience and Remote Sensing, vol. 57, no. 8, pp. 5813-5828. https://doi.org/10.1109/TGRS.2019.2902568

APA

Zhang, H., Li, Y., Jiang, Y., Wang, P., Shen, Q., & Shen, C. (2019). Hyperspectral classification based on lightweight 3D-CNN with transfer learning. IEEE Transactions on Geoscience and Remote Sensing, 57(8), 5813-5828. https://doi.org/10.1109/TGRS.2019.2902568

Vancouver

Zhang H, Li Y, Jiang Y, Wang P, Shen Q, Shen C. Hyperspectral classification based on lightweight 3D-CNN with transfer learning. IEEE Transactions on Geoscience and Remote Sensing. 2019 Aug 1;57(8):5813-5828. https://doi.org/10.1109/TGRS.2019.2902568

Author

Zhang, Haokui ; Li, Ying ; Jiang, Yenan ; Wang, Peng ; Shen, Qiang ; Shen, Chunhua . / Hyperspectral classification based on lightweight 3D-CNN with transfer learning. In: IEEE Transactions on Geoscience and Remote Sensing. 2019 ; Vol. 57, No. 8. pp. 5813-5828.

Bibtex - Download

@article{4efbf28796ee4cba97911d50c0ab8217,
title = "Hyperspectral classification based on lightweight 3D-CNN with transfer learning.",
abstract = "Recently, hyperspectral image (HSI) classification approaches based on deep learning (DL) models have been proposed and shown promising performance. However, due to very limited available training samples and massive model parameters, deep learning methods may suffer from over-fitting. In thispaper, we propose an end-to-end 3D lightweight convolutional neural network (CNN) (abbreviated as 3D-LWNet) for limited samples based HSI classification. Compared with conventional 3D-CNN models, the proposed 3D-LWNet has deeper network structure, less parameters and lower computation cost, while itresults in better classification performance. To further alleviate the small sample problem, we also propose two transfer learning strategies: 1) cross-sensor strategy. We pretrain a 3D model in the source HSI datasets containing a greater number of labeled samples and then transfer it to the target HSI datasets; and2) cross-modal strategy. We pretrain a 3D model in the 2D RGB image datasets containing a large number of samples and then transfer it to the target HSI datasets. In contrast to previous approaches, we do not impose restrictions over the source datasets in that they do not have to be collected by the same senors as the target datasets. Experiments on three public HSI datasets captured by different sensors demonstrate that our model achieves competitive performance for HSI classification compared to several state-of-the-art methods",
keywords = "hyperspectral classification, deep learning, 3D lightweight convolutional network, transfer learning",
author = "Haokui Zhang and Ying Li and Yenan Jiang and Peng Wang and Qiang Shen and Chunhua Shen",
year = "2019",
month = "8",
day = "1",
doi = "10.1109/TGRS.2019.2902568",
language = "English",
volume = "57",
pages = "5813--5828",
journal = "IEEE Transactions on Geoscience and Remote Sensing",
issn = "0196-2892",
publisher = "IEEE Press",
number = "8",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - Hyperspectral classification based on lightweight 3D-CNN with transfer learning.

AU - Zhang, Haokui

AU - Li, Ying

AU - Jiang, Yenan

AU - Wang, Peng

AU - Shen, Qiang

AU - Shen, Chunhua

PY - 2019/8/1

Y1 - 2019/8/1

N2 - Recently, hyperspectral image (HSI) classification approaches based on deep learning (DL) models have been proposed and shown promising performance. However, due to very limited available training samples and massive model parameters, deep learning methods may suffer from over-fitting. In thispaper, we propose an end-to-end 3D lightweight convolutional neural network (CNN) (abbreviated as 3D-LWNet) for limited samples based HSI classification. Compared with conventional 3D-CNN models, the proposed 3D-LWNet has deeper network structure, less parameters and lower computation cost, while itresults in better classification performance. To further alleviate the small sample problem, we also propose two transfer learning strategies: 1) cross-sensor strategy. We pretrain a 3D model in the source HSI datasets containing a greater number of labeled samples and then transfer it to the target HSI datasets; and2) cross-modal strategy. We pretrain a 3D model in the 2D RGB image datasets containing a large number of samples and then transfer it to the target HSI datasets. In contrast to previous approaches, we do not impose restrictions over the source datasets in that they do not have to be collected by the same senors as the target datasets. Experiments on three public HSI datasets captured by different sensors demonstrate that our model achieves competitive performance for HSI classification compared to several state-of-the-art methods

AB - Recently, hyperspectral image (HSI) classification approaches based on deep learning (DL) models have been proposed and shown promising performance. However, due to very limited available training samples and massive model parameters, deep learning methods may suffer from over-fitting. In thispaper, we propose an end-to-end 3D lightweight convolutional neural network (CNN) (abbreviated as 3D-LWNet) for limited samples based HSI classification. Compared with conventional 3D-CNN models, the proposed 3D-LWNet has deeper network structure, less parameters and lower computation cost, while itresults in better classification performance. To further alleviate the small sample problem, we also propose two transfer learning strategies: 1) cross-sensor strategy. We pretrain a 3D model in the source HSI datasets containing a greater number of labeled samples and then transfer it to the target HSI datasets; and2) cross-modal strategy. We pretrain a 3D model in the 2D RGB image datasets containing a large number of samples and then transfer it to the target HSI datasets. In contrast to previous approaches, we do not impose restrictions over the source datasets in that they do not have to be collected by the same senors as the target datasets. Experiments on three public HSI datasets captured by different sensors demonstrate that our model achieves competitive performance for HSI classification compared to several state-of-the-art methods

KW - hyperspectral classification

KW - deep learning

KW - 3D lightweight convolutional network

KW - transfer learning

U2 - 10.1109/TGRS.2019.2902568

DO - 10.1109/TGRS.2019.2902568

M3 - Article

VL - 57

SP - 5813

EP - 5828

JO - IEEE Transactions on Geoscience and Remote Sensing

JF - IEEE Transactions on Geoscience and Remote Sensing

SN - 0196-2892

IS - 8

ER -

Show download statistics
View graph of relations
Citation formats