An Adaptive Capsule Network for Hyperspectral Remote Sensing Classification

Abstract

The capsule network (Caps) is a novel type of neural network that has great potential for the classification of hyperspectral remote sensing. However, the Caps suffers from the issue of gradient vanishing. To solve this problem, a powered activation regularization based adaptive capsule network (PAR-ACaps) was proposed for hyperspectral remote sensing classification, in which an adaptive routing algorithm without iteration was applied to amplify the gradient, and the powered activation regularization method was used to learn the sparser and more discriminative representation. The classification performance of PAR-ACaps was evaluated using two public hyperspectral remote sensing datasets, i.e., the Pavia University (PU) and Salinas (SA) datasets. The average overall classification accuracy (OA) of PAR-ACaps with shallower architecture was measured and compared with those of the benchmarks, including random forest (RF), support vector machine (SVM), 1-dimensional convolutional neural network (1DCNN), two-dimensional convolutional neural network (CNN), three-dimensional convolutional neural network (3DCNN), Caps, and the original adaptive capsule network (ACaps) with comparable network architectures. The OA of PAR-ACaps for PU and SA datasets was 99.51% and 94.52%, respectively, which was higher than those of benchmarks. Moreover, the classification performance of PAR-ACaps with relatively deeper neural architecture (four and six convolutional layers in the feature extraction stage) was also evaluated to demonstrate the effectiveness of gradient amplification. As shown in the experimental results, the classification performance of PAR-ACaps with relatively deeper neural architecture for PU and SA datasets was also superior to 1DCNN, CNN, 3DCNN, Caps, and ACaps with comparable neural architectures. Additionally, the training time consumed by PAR-ACaps was significantly lower than that of Caps. The proposed PAR-ACaps is, therefore, recommended as an effective alternative for hyperspectral remote sensing classification.

Unsupervised multi-branch capsule for hyperspectral and LiDAR classification

“…The spectral methods exploited traditional machine learning models like random forests 7,8 and support vector machines (SVM) 9,10 to extract spectral features from spectral signatures. Subsequently, deep learning such as convolutional neural networks (CNN) [11][12][13] and capsule networks [14][15][16] showed their advantages. The spectralspatial methods extracted spatial and spectral information with CNN.…”

Patch-Free Bilateral Network for Hyperspectral Image Classification Using Limited Samples

“…To this end, the patch level DL classifiers received extensive attention in HSI classification. For example, 3D-CNN [20], 2D-RNN [21], capsule network [22] are used to improve HSI classification accuracy respectively. Meanwhile, the latest deep network architectures such as cascaded recurrent neural networks [23], attention mechanism [24], [25], graph convolutional networks [26], residual learning [27] and densely connected network [28] are used to improve the results of HSI classification.…”