Combining 3D-CNN and Squeeze-and-Excitation Networks for Remote Sensing Sea Ice Image Classification

Abstract

Sea ice is one of the most prominent marine disasters in high latitudes. Remote sensing technology provides an effective means for sea ice detection. Remote sensing sea ice images contain rich spectral and spatial information. However, most traditional methods only focus on spectral information or spatial information, and do not excavate the feature of spectral and spatial simultaneously in remote sensing sea ice images classification. At the same time, the complex correlation characteristics among spectra and small sample problem in sea ice classification also limit the improvement of sea ice classification accuracy. For this issue, this paper proposes a new remote sensing sea ice image classification method based on squeeze-and-excitation (SE) network, convolutional neural network (CNN), and support vector machines (SVMs). The proposed method designs 3D-CNN deep network so as to fully exploit the spatial-spectrum features of remote sensing sea ice images and integrates SE-Block into 3D-CNN in-depth network in order to distinguish the contributions of different spectra to sea ice classification. According to the different contributions of spectral features, the weight of each spectral feature is optimized by fusing SE-Block in order to further enhance the sample quality. Finally, information-rich and representative samples are chosen by combining the idea of active learning and input into SVM classifier, and this achieves superior classification accuracy of remote sensing sea ice images with small samples. In order to verify the effectiveness of the proposed method, we conducted experiments on three different data from Baffin Bay, Bohai Bay, and Liaodong Bay. The experimental results show that compared with other classical classification methods, the proposed method comprehensively considers the correlation among spectral features and the small samples problems and deeply excavates the spatial-spectrum characteristics of sea ice and achieves better classification performance, which can be effectively applied to remote sensing sea ice image classification.

Understanding Dilated Mathematical Relationship between Image Features and the Convolutional Neural Network’s Learnt Parameters

“…This excitation operation is applied to generate the model’s weights for each feature channel with parameters, where these parameters learn to model the effective correlation between the feature channels. Therefore, the excitation operation must be flexible to learn the non-linear interactions and the non-mutually exclusive relationship between channels to capture the channel wise dependency [ 32 ]. Hence, during the development of deep learning technology, it was proved that the updates on a model’s excitation weights are mainly related to training the samples’ extracted feature values, assuming an overall mathematical relationship between them.…”

Earthquake Damage Assessment Based on Deep Learning Method Using VHR Images

“…Figure 3 shows the general view of the deep learning networks. Unlike deep learning, machine learning extracts features by itself, and they need to identify the characteristics and feature engineering [24][25][26].…”

ME-CapsNet: A Multi-Enhanced Capsule Networks with Routing Mechanism

“…A lot of customization to SENets have also been done, including Yue Cao et al [13] performing a unification of NLnet and SENet resulting in a Global Context (GC) network. A lot of SENets applications has been established including some works by Qiu et al [14], Shunjun Wei et al [15], and Han et al [16] where fusion of SENets were done with various architectures of CNN for fish image classifications, PRI modulation recognition, sea ice image sensing respectively.…”