4.7 Article

CNN-Enhanced Graph Convolutional Network With Pixel- and Superpixel-Level Feature Fusion for Hyperspectral Image Classification

Journal

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
Volume 59, Issue 10, Pages 8657-8671

Publisher

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TGRS.2020.3037361

Keywords

Feature extraction; Convolution; Kernel; Decoding; Training data; Deep learning; Computational modeling; Convolutional neural network (CNN); feature fusion; graph convolutional network (GCN); graph encoder and decoder; hyperspectral image (HSI) classification

Funding

  1. National Natural Science Foundation of China [61871226, 62001226, 61571230]
  2. Jiangsu Provincial Social Developing Project [BE2018727]
  3. Fundamental Research Funds for the Central Universities [30918011104, 30920021134]
  4. National Major Research Plan of China [2016YFF0103604]

Ask authors/readers for more resources

This paper proposes a heterogeneous deep network called CEGCN, which integrates CNN and GCN branches for feature learning on different scales of image regions to generate complementary spectral-spatial features. By integrating the graph encoding process into the network and learning edge weights from training data, it promotes node feature learning and makes the graph more adaptive to HSI content.
Recently, the graph convolutional network (GCN) has drawn increasing attention in the hyperspectral image (HSI) classification. Compared with the convolutional neural network (CNN) with fixed square kernels, GCN can explicitly utilize the correlation between adjacent land covers and conduct flexible convolution on arbitrarily irregular image regions; hence, the HSI spatial contextual structure can be better modeled. However, to reduce the computational complexity and promote the semantic structure learning of land covers, GCN usually works on superpixel-based nodes rather than pixel-based nodes; thus, the pixel-level spectral-spatial features cannot be captured. To fully leverage the advantages of the CNN and GCN, we propose a heterogeneous deep network called CNN-enhanced GCN (CEGCN), in which CNN and GCN branches perform feature learning on small-scale regular regions and large-scale irregular regions, and generate complementary spectral-spatial features at pixel and superpixel levels, respectively. To alleviate the structural incompatibility of the data representation between the Euclidean data-oriented CNN and non-Euclidean data-oriented GCN, we propose the graph encoder and decoder to propagate features between image pixels and graph nodes, thus enabling the CNN and GCN to collaborate in a single network. In contrast to other GCN-based methods that encode HSI into a graph during preprocessing, we integrate the graph encoding process into the network and learn edge weights from training data, which can promote the node feature learning and make the graph more adaptive to HSI content. Extensive experiments on three data sets demonstrate that the proposed CEGCN is both qualitatively and quantitatively competitive compared with other state-of-the-art methods.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available