Abstract:Deep learning based methods have seen a massive rise in popularity for hyperspectral image classification over the past few years. However, the success of deep learning is attributed greatly to numerous labeled samples. It is still very challenging to use only a few labeled samples to train deep learning models to reach a high classification accuracy. An active deep-learning framework trained by an end-to-end manner is, therefore, proposed by this paper in order to minimize the hyperspectral image classification costs. First, a deep densely connected convolutional network is considered for hyperspectral image classification. Different from the traditional active learning methods, an additional network is added to the designed deep densely connected convolutional network to predict the loss of input samples. Then, the additional network could be used to suggest unlabeled samples that the deep densely connected convolutional network is more likely to produce a wrong label. Note that the additional network uses the intermediate features of the deep densely connected convolutional network as input. Therefore, the proposed method is an end-to-end framework. Subsequently, a few of the selected samples are labelled manually and added to the training samples. The deep densely connected convolutional network is therefore trained using the new training set. Finally, the steps above are repeated to train the whole framework iteratively. Extensive experiments illustrates that the method proposed could reach a high accuracy in classification after selecting just a few samples.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to train a deep - learning model to achieve high classification accuracy in hyperspectral image classification when there are only a small number of labeled samples. Specifically, the paper addresses the following two main challenges: 1. **High - Dimensional Data**: Hyperspectral images (HSI) have high - dimensional characteristics, which makes it difficult to directly use traditional deep - learning methods for classification because a large amount of feature information needs to be processed. 2. **Insufficient Labeled Samples**: Obtaining a large number of labeled samples is costly and time - consuming, which limits the application of deep - learning models in hyperspectral image classification. To meet these challenges, the paper proposes an Active Deep Learning Framework, which combines the Deep Densely Connected Convolutional Network (DDCN) and the Loss Prediction Model. The specific methods are as follows: 1. **Deep Densely Connected Convolutional Network**: - A deep densely connected convolutional network based on the classic DenseNet121 is designed for hyperspectral image classification. - By increasing the number of network layers, the classification performance of the model is improved. 2. **Loss Prediction Model**: - An additional network is introduced to predict the loss values of input samples, thereby identifying those samples that are more likely to be misclassified. - These selected samples will be manually labeled and added to the training set to gradually improve the classification performance of the model. 3. **Iterative Training**: - In the initial stage, a small number of samples are randomly selected for manual labeling to form an initial training set. - The DDCN and the loss prediction model are trained using the initial training set. - According to the output of the loss prediction model, the samples that are most likely to be misclassified are selected for labeling and added to the training set. - The above process is repeated to gradually increase the size of the training set until satisfactory classification accuracy is achieved. Through this active learning strategy, the paper aims to reduce the labeling cost while improving the accuracy of hyperspectral image classification. The experimental results show that this method can achieve high classification accuracy on multiple hyperspectral image datasets, especially when the number of labeled samples is limited.

Active Deep Densely Connected Convolutional Network for Hyperspectral Image Classification

Deep Convolutional Neural Networks for Hyperspectral Image Classification

Active Deep Learning for Classification of Hyperspectral Images

3D Convolutional Siamese Network for Few-Shot Hyperspectral Classification

Semi-supervised Deep Learning for Hyperspectral Image Classification

Convolutional Neural Network Based Classification for Hyperspectral Data

Hyperspectral Image Classification Based on Deep Separable Residual Attention Network

A Lightweight Convolutional Neural Network for Hyperspectral Image Classification

Semi-Active Convolutional Neural Networks for Hyperspectral Image Classification

Deep Multilayer Fusion Dense Network for Hyperspectral Image Classification

Three-dimensional Densely Connected Convolutional Network for Hyperspectral Remote Sensing Image Classification

Deep Prototypical Networks With Hybrid Residual Attention for Hyperspectral Image Classification

A Fast Dense Spectral-Spatial Convolution Network Framework for Hyperspectral Images Classification

Hyperspectral Image Classification With Deep Learning Models

Hyperspectral Classification Based on Spectral-Spatial Convolutional Neural Networks

Hyperspectral Image Classification Based on Deep Deconvolution Network with Skip Architecture

Spectral–spatial multi-feature-based deep learning for hyperspectral remote sensing image classification

Active Transfer Learning Network: A Unified Deep Joint Spectral–Spatial Feature Learning Model for Hyperspectral Image Classification

An Efficient Hyperspectral Image Classification Method for Limited Training Data

A hybrid approach consisting of 3D depthwise separable convolution and depthwise squeeze-and-excitation network for hyperspectral image classification

Deep Convolutional Networks with Superpixel Segmentation for Hyperspectral Image Classification