Figure 2 - available via license: CC BY
Content may be subject to copyright.
Source publication
- [...]
The classification of hyperspectral data using deep learning methods can obtain better results than the previous shallow classifiers, but deep learning algorithms have some limitations. These algorithms require a large amount of data to train the network, while also needing a certain amount of labeled data to fine-tune the network. In this paper, w...
Contexts in source publication
Context 1
... restricted Boltzmann machine is a special topology of the Boltzmann machine (BM). The topology of the Boltzmann machine network is shown in Figure 2 [24]. We use the source domain data for the deep learning network training, and we use the target domain's limited labeled data for fine-tuning the network. ...
Similar publications
Fault diagnosis plays a very important role in ensuring the safe and reliable operations of machines. Currently, the deep learning-based fault diagnosis is attracting increasing attention. However, fault diagnosis under variable working conditions has been a significant challenge due to the domain discrepancy problem. This problem is also unavoidab...
Citations
... The classification of these objects, which entails stratifying the pixels in the hyperspectral image (HSI), is an essential process in various fields, such as remote sensing, computer vision, environmental monitoring, and resource management [2]. Various deep learning models have been utilized for HSI classification [3], including both supervised and semi-supervised, such as autoencoders (AE) [4], convolutional neural networks (CNN) [5], long short-term memory (LSTM) [6], transfer learning (TL) [7], deep belief network (DBN) [8], have been introduced as these models have data-driven feature learning capabilities. Among the aforementioned classes of models, the supervised networks [9][10][11] have evolved to achieve almost perfect classification accuracy. ...
The classification of hyperspectral images (HSI) into categories that correlate to various land cover sorts such as water bodies, agriculture and urban areas, has gained significant attention in research due to its wide range of applications in fields, such as remote sensing, computer vision, and more. Supervised deep learning networks have demonstrated exceptional performance in HSI classification, capitalizing on their capacity for end-to-end optimization and leveraging their strong potential for nonlinear modeling. However, labelling HSIs, on the other hand, necessitates extensive domain knowledge and is a time-consuming and labour-intensive exercise. To address this issue, the proposed work introduces a novel semi-supervised network constructed with an autoencoder, Siamese action, and attention layers that achieves excellent classification accuracy with labelled limited samples. The proposed convolutional autoencoder is trained using the mass amount of unlabelled data to learn the refinement representation referred to as 3D-CAE. The added Siamese network improves the feature separability between different categories and attention layers improve classification by focusing on discriminative information and neglecting the unimportant bands. The efficacy of the proposed model’s performance was assessed by training and testing on both same-domain as well as cross-domain data and found to achieve 91.3 and 93.6 for Indian Pines and Salinas, respectively.
... Although the classification effect of deep learning is better than that of shallow classifier, it needs a certain number of labeled samples to build the model. Li et al. [85] proposed a hyperspectral data processing method based on transfer learning and deep learning. The hyperspectral data sets similar to the target data were used to pre train the deep learning network, found the common features and established the classification model. ...
... This leverages the Markov property of the images to separate images with class tags and train the CNN on random band samples selected from the datasets. The paper by Ke Li et al. [13] explores another transfer learning approach for hyperspectral image classification using deep belief networks wherein a limited Boltzmann machine Network is trained on the source domain data and its first few layers are extracted to be used for the target domain network. The target domain network is fine tuned further and used for classification of images in the target domain. ...
A Hyperspectral image contains much more number of channels as compared to a RGB image, hence containing more information about entities within the image. The convolutional neural network (CNN) and the Multi-Layer Perceptron (MLP) have been proven to be an effective method of image classification. However, they suffer from the issues of long training time and requirement of large amounts of the labeled data, to achieve the expected outcome. These issues become more complex while dealing with hyperspectral images. To decrease the training time and reduce the dependence on large labeled dataset, we propose using the method of transfer learning. The hyperspectral dataset is preprocessed to a lower dimension using PCA, then deep learning models are applied to it for the purpose of classification. The features learned by this model are then used by the transfer learning model to solve a new classification problem on an unseen dataset. A detailed comparison of CNN and multiple MLP architectural models is performed, to determine an optimum architecture that suits best the objective. The results show that the scaling of layers not always leads to increase in accuracy but often leads to overfitting, and also an increase in the training time.The training time is reduced to greater extent by applying the transfer learning approach rather than just approaching the problem by directly training a new model on large datasets, without much affecting the accuracy.
... Shutao Li et al provided a detailed comparative study of the specific deep learning frameworks used for hyperspectral identification including Deep Belief Networks(DBN), Generative Adversarial Networks(GAN's), Recurrent Neural Networks(RNNs) and Convolutionary Neural Networks(CNNs) and numerous variants such as the Gabor-CNN, S-CNN and RESNET. [29] The paper by Ke Li et al. [34] explores another transfer learning approach for hyper-spectral image classification using deep belief networks wherein a limited Boltzmann machine Network is trained on the source domain data and its first few layers are extracted to be used for the target domain network. The target domain network is fine tuned further and used for classification of images in the target domain. ...
Hyper-spectral images are images captured from a satellite that gives spatial and spectral information of specific region.A Hyper-spectral image contains much more number of channels as compared to a RGB image, hence containing more information about entities within the image. It makes them well suited for the classification of objects in a snap. In the past years, the efficiency of hyper-spectral image recognition has increased significantly with deep learning. The Convolution Neural Network(CNN) and Multi-Layer Perceptron(MLP) has demonstrated to be an excellent process of classifying images. However, they suffer from the issues of long training time and requirement of large amounts of the labeled data, to achieve the expected outcome. These issues become more complex while dealing with hyper-spectral images. To decrease the training time and reduce the dependence on large labeled data-set, we propose using the method of transfer learning.The features learned by CNN and MLP models are then used by the transfer learning model to solve a new classification problem on an unseen dataset. A detailed comparison of CNN and multiple MLP architectural models is performed, to determine an optimum architecture that suits best the objective. The results show that the scaling of layers not always leads to increase in accuracy but often leads to over-fitting, and also an increase in the training time.The training time is reduced to greater extent by applying the transfer learning approach rather than just approaching the problem by directly training a new model on large data-sets, without much affecting the accuracy.
... Other approaches include using labels at a coarser granularity for the target domain (e.g. image-level labels) as weak supervision [21] and learning latent representations shared between source and target do-mains to help in adaptation [26,29]. ...
The increasing availability of high-resolution satellite imagery has enabled the use of machine learning to support land-cover measurement and inform policy-making. However, labelling satellite images is expensive and is available for only some locations. This prompts the use of transfer learning to adapt models from data-rich locations to others. Given the potential for high-impact applications of satellite imagery across geographies, a systematic assessment of transfer learning implications is warranted. In this work, we consider the task of land-cover segmentation and study the fairness implications of transferring models across locations. We leverage a large satellite image segmentation benchmark with 5987 images from 18 districts (9 urban and 9 rural). Via fairness metrics we quantify disparities in model performance along two axes -- across urban-rural locations and across land-cover classes. Findings show that state-of-the-art models have better overall accuracy in rural areas compared to urban areas, through unsupervised domain adaptation methods transfer learning better to urban versus rural areas and enlarge fairness gaps. In analysis of reasons for these findings, we show that raw satellite images are overall more dissimilar between source and target districts for rural than for urban locations. This work highlights the need to conduct fairness analysis for satellite imagery segmentation models and motivates the development of methods for fair transfer learning in order not to introduce disparities between places, particularly urban and rural locations.
... Although the classification effect of deep learning is better than that of shallow classifier, it needs a certain number of labeled samples to build the model. Li et al. [85] proposed a hyperspectral data processing method based on transfer learning and deep learning. The hyperspectral data sets similar to the target data were used to pre train the deep learning network, found the common features and established the classification model. ...
Hyperspectral image (HSI) contains rich spatial and spectral information, which has been widely used in resource exploration, ecological environment monitoring, land cover classification and target recognition. However, the nonlinearity of HSI data and the strong correlation between bands also bring difficulties and challenges to HSI application. In particular, the limited available hyperspectral training samples will lead to the classification accuracy cannot be improved. Therefore, making full use of the advantages of HSI data, through algorithms and strategies to solve the limited training samples, high-dimensional HSI data and effective classification method, so as to improve the classification accuracy. This paper reviews the research results of the feature extraction methods and classification methods of HSI classification in recent years. In addition, this paper expounds five kinds of small sample strategies, and solves the problem of small sample in HSI classification from different angles. Small sample strategy will be the focus of HSI classification research in the future. To solve the problem of small sample classification can greatly promote the application of HSI.
... Transfer learning based on these cross-domain invariant features can effectively reduce the differences between the source domain and the target domain, and align the distributions of the source domain and the target domain. Based on this idea, the method based on deep transfer learning [15][16][17][18][19][20] has been successfully applied in various tasks. Among them, the domain adaptive method, as a special form of transfer learning, has been widely used in fault diagnosis problems, and the learned diagnosis knowledge can be well extended to different mechanical conditions. ...
Deep learning-based mechanical fault diagnosis method has made great achievements. A high-performance neural network model requires sufficient labelled data for training to obtain accurate classification results. Desired results mainly depends on assumption that training and testing data are collected under the same working conditions, environment and operating conditions, where the data have the same probability distribution. However, in the practical scenarios, training data and the testing data follow different distributions to some degree, and the newly collected testing data are usually unlabeled. In order to solve the problems above, a model based on transfer learning and domain adaptation is proposed to achieve efficient fault diagnosis under different data distributions. The proposed framework adapts the features extracted by multiple dynamic convolutional layers, and creatively utilizes correlation alignment(CORAL) to perform a non-linear transformation to align the second-order statistics of the two distributions for fault diagnosis, which greatly improves the accuracy of fault classification in the target domain under unlabeled data. Finally, experimental verifications have been carried out among two different datasets.
... However, such methods usually exhibit two common issues: (1) How to develop Their experimental results show that the pre-training network is beneficial to improve the classification accuracy, and the datasets of different HSI sensors are useful for the pre-training. Ke Li et al. [22] have proposed a new hyperspectral classification method based on transfer learning and the deep learning method. They used a hyperspectral dataset that is similar to the target dataset to pre-training the deep learning network. ...
Since hyperspectral images (HSI) captured by different sensors often contain different number of bands, but most of the convolutional neural networks (CNN) require a fixed-size input, the generalization capability of deep CNNs to use heterogeneous input to achieve better classification performance has become a research focus. For classification tasks with limited labeled samples, the training strategy of feeding CNNs with sample-pairs instead of single sample has proven to be an efficient approach. Following this strategy, we propose a Siamese CNN with three-dimensional (3D) adaptive spatial-spectral pyramid pooling (ASSP) layer, called ASSP-SCNN, that takes as input 3D sample-pair with varying size and can easily be transferred to another HSI dataset regardless of the number of spectral bands. The 3D ASSP layer can also extract different levels of 3D information to improve the classification performance of the equipped CNN. To evaluate the classification and generalization performance of ASSP-SCNN, our experiments consist of two parts: the experiments of ASSP-SCNN without pre-training and the experiments of ASSP-SCNN-based transfer learning framework. Experimental results on three HSI datasets demonstrate that both ASSP-SCNN without pre-training and transfer learning based on ASSP-SCNN achieve higher classification accuracies than several state-of-the-art CNN-based methods. Moreover, we also compare the performance of ASSP-SCNN on different transfer learning tasks, which further verifies that ASSP-SCNN has a strong generalization capability.
... With the development of deep learning technology, more deep learning models are applied to spectral data processing [16][17][18]. The sparse autoencoder (SAE) is an unsupervised deep learning algorithm used in the reduction of data dimension and extraction of features [19,20]. ...
The soluble solids content (SSC) affects the flavor of green plums and is an important parameter during processing. In recent years, the hyperspectral technology has been widely used in the nondestructive testing of fruit ingredients. However, the prediction accuracy of most models can hardly be improved further. The rapid development of deep learning technology has established the foundation for the improvement of building models. A new hyperspectral imaging system aimed at measuring the green plum SSC is developed, and a sparse autoencoder (SAE)–partial least squares regression (PLSR) model is combined to further improve the accuracy of component prediction. The results of the experiment show that the SAE–PLSR model, which has a correlation coefficient of 0.938 and root mean square error of 0.654 for the prediction set, can achieve better performance for the SSC prediction of green plums than the three traditional methods. In this paper, integration approaches have combined three different pretreatment methods with PLSR to predict the SSC in green plums. The SAE–PLSR model has shown good prediction performance, indicating that the proposed SAE–PLSR model can effectively detect the SSC in green plums.
... It consists of multiple restricted Boltzmann machines (RBMs) and a back propagation (BP) neural network, allowing processing large amounts of data. In recent decades, the DBN has been successfully applied to fault diagnosis, wind speed forecasting, breast cancer classification, and so on [11][12][13][14][15][16]. Due to its advantage in prediction accuracy, in this paper, we introduce the DBN to conduct the PM2.5 concentration prediction. ...
Accurate PM2.5 concentration prediction is crucial for protecting public health and improving air quality. As a popular deep learning model, deep belief network (DBN) for PM2.5 concentration prediction has received increasing attention due to its effectiveness. However, the DBN structure parameters that have a significant impact on prediction accuracy and computation time are hard to be determined. To address this issue, a modified grey wolf optimization (MGWO) algorithm is proposed to optimize the DBN structure parameters containing number of hidden nodes, learning rate, and momentum coefficient. The methodology modifies the basic grey wolf optimization (GWO) algorithm using the nonlinear convergence and position update strategies, and then utilizes the training error of the DBN to calculate the fitness function of the MGWO algorithm. Through the multiple iterations, the optimal structure parameters are obtained, and a suitable predictor is finally generated. The proposed prediction model is validated on a real application case. Compared with the other prediction models, experimental results show that the proposed model has a simpler structure but higher prediction accuracy.
