Learning and Transferring Deep Joint Spectral–Spatial Features for Hyperspectral Classification

By Jingxiang Yang, Yongqiang Zhao, Jonathan Cheung-Wai Chan. IEEE Transactions on Geoscience and Remote Sensing, 2017

437 citations recorded by the source corpus.

Subfields: Atmospheric Science, Biomedical Engineering, Media Technology

View Paper: https://www.semanticscholar.org/paper/699a373cd3d302f0218d8213ca2fb7b543fc1e2d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Atmospheric Science

1. Shutao Li
2. Jon Atli Benediktsson
3. Pedram Ghamisi
4. Liangpei Zhang
5. Xudong Kang
6. Leyuan Fang
7. Jocelyn Chanussot
8. Yushi Chen

In Biomedical Engineering

1. Shutao Li
2. Xudong Kang
3. Leyuan Fang
4. Pedram Ghamisi
5. Jon Atli Benediktsson
6. Liangpei Zhang
7. Lorenzo Bruzzone
8. Danfeng Hong

In Media Technology

1. Shutao Li
2. Jocelyn Chanussot
3. Jon Atli Benediktsson
4. Liangpei Zhang
5. Pedram Ghamisi
6. Leyuan Fang
7. Antonio J. Plaza
8. Danfeng Hong

Citations in the corpus

• Attention-Based Adaptive Spectral–Spatial Kernel ResNet for Hyperspectral Image Classification (2020, background): “[28] proposed a two-branch architecture to extract the joint spectral–spatial features for classification.”
• Cascaded Recurrent Neural Networks for Hyperspectral Image Classification (2017, methodology): “In [41]–[43], 1-D CNN or autoencoder was used to learn”
• Spectral–Spatial Feature Tokenization Transformer for Hyperspectral Image Classification (2022, methodology): “In [39], to extract spatial and spectral features, a dual-branch network structure was developed by combining 2-D-and 1-D-CNN technologies.”
• CNN-Enhanced Graph Convolutional Network With Pixel- and Superpixel-Level Feature Fusion for Hyperspectral Image Classification (2020, methodology): “While in [26], the two-channel CNN (TCCNN) was proposed to integrate 1-”
• Feature Extraction for Hyperspectral Imagery: The Evolution From Shallow to Deep: Overview and Toolbox (2020, methodology): “In [108] and [114], a parallel structure was employed to extract spectral-spatial features.”
• Spectral-Spatial Attention Networks for Hyperspectral Image Classification (2019, methodology): “[15] designed a two-CNN model to learn the spectral features and spatial features jointly.”
• Hyperspectral Image Classification—Traditional to Deep Models: A Survey for Future Prospects (2021, background): “For instance, [225] pertained a two branch spectralspatial CNN model with an ample amount of training data from”
• Deep Learning for Hyperspectral Image Classification: An Overview (2019, methodology): “For instance, in [107] and [108], a deep CNN with two-branch architecture was proposed to extract the joint spectral–spatial features from HSIs.”
• A Spectral-Spatial-Dependent Global Learning Framework for Insufficient and Imbalanced Hyperspectral Image Classification (2021, methodology): “Then the bottom layers of the pretrained network were transferred to the target network as initialization, and the top layers were randomly initialized [42].”
• Deep Learning Meets Hyperspectral Image Analysis: A Multidisciplinary Review (2019, background): “Yang et al. [152] used a two stream spatial–spectral network to perform transfer-learning, by fine-tuning only final layers, producing an improvement with respect to excluding the transfer-learning part.”
• A survey: Deep learning for hyperspectral image classification with few labeled samples (2021, background): “The later transfer learning models based on fine-tuning basically follow that architecture [85, 86, 87, 88].”