Going Deeper With Contextual CNN for Hyperspectral Image Classification

Abstract

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.

DenseNet-Based Depth-Width Double Reinforced Deep Learning Neural Network for High-Resolution Remote Sensing Image Per-Pixel Classification

“…When using the multi-scale filters for feature extraction, we choose a method that differs from the one-time convolution used in many current studies [17], which focuses on using the multi-scale filters to convolve the original input data, then integrating the feature maps generated by the process and inputting them into the following network with convolution based on one fixed-scale filter. We utilize multi-scale filters to extract hierarchical features several times.…”

“…To test the effectiveness and practicability of the proposed method, we introduced five kinds of networks similar to the proposed method: DNN [47], URDNN [48], contextual deep CNN [17], two-stream neural network [26] and SAE + SVM [49]. DNN uses deconvolution to realize an end-to-end, pixel-to-pixel classification.…”

When Low Rank Representation Based Hyperspectral Imagery Classification Meets Segmented Stacked Denoising Auto-Encoder Based Spatial-Spectral Feature

“…A deep neural network (DNN) is a kind of machine learning method; the basic idea is to build a neural network model containing multiple hidden layers, which needs a large amount of training data to train the network model. There are already many supervised DNN methods [12,[14][15][16][17][22][23][24][25][26]. In [22], deep convolutional neural networks (DCNNs) are employed to classify HSIs directly in the spectral domain.…”

An analysis of Feature extraction and Classification Algorithms for Dangerous Object Detection

“…Alexnet is a deep CNN created by Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton which is trained with over a million images from 1000 different classes [20]. Alexnet consists of five convolutional layers and three fully connected layers [21]. Figure 1 shows all the steps of the experiment.…”