Contextual deep CNN based hyperspectral classification

Lee, Hyungtae; Kwon, Heesung

doi:10.1109/igarss.2016.7729859

Cited by 100 publications

(81 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It can also be observed that the proposed method performs better or equally well when compared with other state-ofthe art methods. More specifically, the proposed method outperforms [18] and [19] by 9.58% and 5.61% on PaviaU dataset with 4.4% training data. Similarly, our model Table 5.…”

Section: Resultsmentioning

confidence: 92%

“…The proposed method has been trained and evaluated over PaviaC and PaviaU datasets. The performance is compared with recent state-of-the-art approaches [18,19] in terms of classification accuracy. We trained our model with different percentage of training data (4.4%, 5%, 9%, 15%).…”

Section: Resultsmentioning

confidence: 99%

“…It involves removal of unwanted, irrelevant bands from each pixel of the image. The existing approaches for hyperspectral image classification (HSIC) can be grouped into traditional [3-7, 16, 17] and deep learning-based methods [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35]. Deep learning models have been much more effective in HSIC as compared to the traditional approaches.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

3D CNN with Localized Residual Connections for Hyperspectral Image Classification

Dwivedi

Mandal

Yadav

et al. 2020

Communications in Computer and Information Science

View full text Add to dashboard Cite

In this paper we propose a novel 3D CNN network with localized residual connections for hyperspectral image classification. Our work chalks a comparative study with the existing methods employed for abstracting deeper features and propose a model which incorporates residual features from multiple stages in the network. The proposed architecture processes individual spatiospectral feature rich cubes from hyperspectral images through 3D convolutional layers. The residual connections result in improved performance due to assimilation of both low-level and high-level features. We conduct experiments over Pavia University and Pavia Center dataset for performance analysis. We compare our method with two recent state-of-the-art methods for hyperspectral image classification method. The proposed network outperforms the existing approaches by a good margin.

show abstract

Section: Resultsmentioning

confidence: 92%

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

3D CNN with Localized Residual Connections for Hyperspectral Image Classification

Dwivedi

Mandal

Yadav

et al. 2020

Communications in Computer and Information Science

View full text Add to dashboard Cite

show abstract

“…Yue et al extracted image features from 46,697 hyperspectral remote sensing images containing 103 bands using a CNN approach from a pixel matrix with spectral and spatial features, and classified the features via logical regression [4]. Lee and Kwon used two convolution kernel templates of different sizes to extract a variety of remote sensing image features from 8,504 samples containing 220 bands and removed the fully-connected layer for classification purposes [5]. All of the above CNNs required numerous, manually labelled samples; as a result, personal experience likely significantly impacted the classification accuracy; furthermore, the dimensionality of the original remote sensing images was high and included considerable redundant information, which affects the efficiency of network training and impedes network learning.…”

Section: Related Workmentioning

confidence: 99%

“…CNNs are typically applied in remote sensing classification efforts in two ways: pixel-based classifications [3][4][5] and scene classifications [6,7]. While there are only a few studies that address object-oriented remote sensing classification based on CNNs, the first-place winner of Dstl's Satellite Imagery competition, Kaggle, proposed a new, improved U-Net model to identify and label significant objects using satellite imagery.…”

Section: Introductionmentioning

confidence: 99%

Application of a Hybrid Model Based on a Convolutional Auto-Encoder and Convolutional Neural Network in Object-Oriented Remote Sensing Classification

2018

View full text Add to dashboard Cite

Variation in the format and classification requirements for remote sensing data makes establishing a standard remote sensing sample dataset difficult. As a result, few remote sensing deep neural network models have been widely accepted. We propose a hybrid deep neural network model based on a convolutional auto-encoder and a complementary convolutional neural network to solve this problem. The convolutional auto-encoder supports feature extraction and data dimension reduction of remote sensing data. The extracted features are input into the convolutional neural network and subsequently classified. Experimental results show that in the proposed model, the classification accuracy increases from 0.916 to 0.944, compared to a traditional convolutional neural network model; furthermore, the number of training runs is reduced from 40,000 to 22,000, and the number of labelled samples can be reduced by more than half, all while ensuring a classification accuracy of no less than 0.9, which suggests the effectiveness and feasibility of the proposed model.

show abstract

Deep learning for remote sensing image classification: A survey

Zhang

Xue

et al. 2018

WIREs Data Min & Knowl

295

146

View full text Add to dashboard Cite

Remote sensing (RS) image classification plays an important role in the earth observation technology using RS data, having been widely exploited in both military and civil fields. However, due to the characteristics of RS data such as high dimensionality and relatively small amounts of labeled samples available, performing RS image classification faces great scientific and practical challenges. In recent years, as new deep learning (DL) techniques emerge, approaches to RS image classification with DL have achieved significant breakthroughs, offering novel opportunities for the research and development of RS image classification. In this paper, a brief overview of typical DL models is presented first. This is followed by a systematic review of pixel‐wise and scene‐wise RS image classification approaches that are based on the use of DL. A comparative analysis regarding the performances of typical DL‐based RS methods is also provided. Finally, the challenges and potential directions for further research are discussed. This article is categorized under: Application Areas > Science and Technology Technologies > Classification

show abstract

Contextual deep CNN based hyperspectral classification

Cited by 100 publications

References 4 publications

3D CNN with Localized Residual Connections for Hyperspectral Image Classification

3D CNN with Localized Residual Connections for Hyperspectral Image Classification

Application of a Hybrid Model Based on a Convolutional Auto-Encoder and Convolutional Neural Network in Object-Oriented Remote Sensing Classification

Deep learning for remote sensing image classification: A survey

Contact Info

Product

Resources

About