Going Deeper With Contextual CNN for Hyperspectral Image Classification

By Hyungtae Lee, Heesung Kwon. 2016

930 citations recorded by the source corpus.

In this paper, we describe a novel deep convolutional neural network (CNN) that is deeper and wider than other existing deep networks for hyperspectral image classification. Unlike current state-of-the-art approaches in CNN-based hyperspectral image classification, the proposed network, called contextual deep CNN, can optimally explore local contextual interactions by jointly exploiting local spatio-spectral relationships of neighboring individual pixel vectors. The joint exploitation of the spatio-spectral information is achieved by a multi-scale convolutional filter bank used as an initial component of the proposed CNN pipeline. The initial spatial and spectral feature maps obtained from the multi-scale filter bank are then combined together to form a joint spatio-spectral feature map. The joint feature map representing rich spectral and spatial properties of the hyperspectral image is then fed through a fully convolutional network that eventually predicts the corresponding label of each pixel vector. The proposed approach is tested on three benchmark data sets: the Indian Pines data set, the Salinas data set, and the University of Pavia data set. Performance comparison shows enhanced classification performance of the proposed approach over the current state-of-the-art on the three data sets.

Subfields: Computer Vision and Pattern Recognition, Media Technology

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

ScaleNet: An Unsupervised Representation Learning Method for Limited Information (2023, methodology): “This experiment clearly shows the adjustment of the ScaleNet to understand and detect influential features with limited information, that will be crucial for many fields with a lack of data [13,33,42].”
Deep learning classifiers for hyperspectral imaging: A review (2019, background): “It has been successfully applied by (Lee et al., 2016; Lee and Kwon, 2017).”
Deep Learning for Classification of Hyperspectral Data: A Comparative Review (2019, positive): “Indeed, many models from the literature are implemented in the toolbox, including 1D CNNs for spectral classification (Figure 7, with 1D convolutions in pink, pooling in orange, and fully connected layers in blue) and the most recent 2D and 3D CNNs as suggested in [49], [63], [68], and [71] (see Fig”
Multiscale Dynamic Graph Convolutional Network for Hyperspectral Image Classification (2019, background): “Furthermore, Lee et al. [29] designed a novel contextual deep CNN, which is able to optimally explore contextual interactions by exploiting local spectral-spatial relationship among spatially neighboring pixels.”
Diverse Region-Based CNN for Hyperspectral Image Classification (2018, methodology): “Different from commonly-used CNN models that apply a sliding window with a specific scale to extract features [36], [39], [40], we present a diverse region-based deep CNN model (denoted as DR-CNN) in this paper.”
Deep Learning for Land Use and Land Cover Classification Based on Hyperspectral and Multispectral Earth Observation Data: A Review (2020, methodology): “In some models the residual blocks are used in a customised network [73,74,123] and in many others, the well-known ResNet models are directly employed to perform the land cover classification of MSI and HSI [120–122,124,125].”
Classification of Hyperspectral Image Based on Double-Branch Dual-Attention Mechanism Network (2019, methodology): “CDCNN: The architecture of the CDCNN is rendered in [56], which is based on 2D CNN and ResNet.”
A Spectral-Spatial-Dependent Global Learning Framework for Insufficient and Imbalanced Hyperspectral Image Classification (2021, background): “patches and classify each patch into one corresponding label rather than directly classifying the whole image [34], [35].”
Deep Learning for Hyperspectral Image Classification: An Overview (2019, methodology): “In addition, research works [87], [88], [93], [97], [98] introduce RL [100] to build a very deep network with the purpose of extracting more discriminative features for HSI classification.”
Solving Current Limitations of Deep Learning Based Approaches for Plant Disease Detection (2019, background): “Convolutional neural networks (CNNs) have performed well in crop classiﬁcation tasks [5], fruit counting, yield prediction [6], disease detection [7–9], and vision tasks in general [9,10].”
Attention-Based Adaptive Spectral–Spatial Kernel ResNet for Hyperspectral Image Classification (2020, methodology): “We compare classical machine learning and representative deep learning methods available in [63] 2 : random forest (RF) [64], multinomial logistic regression (MLR) [65], support vector machine (SVM) with radial basis function [1], gated recurrent unit (GRU) [66], long short term memory (LSTM) [67], ”
A survey: Deep learning for hyperspectral image classification with few labeled samples (2021, methodology): “In the early works of applying CNNs for HSI classification, two-dimensional convolution was the most widely used method, which is mainly employed to extract spatial texture information [47, 28, 48], but the redundant bands greatly”