Search

Deep learning methods have shown considerable potential for hyperspectral image (HSI) classification, which can achieve high accuracy compared with traditional methods. However, they often need a large number of training samples and have a lot of parameters and high computational overhead. To solve these problems, this article proposes new network architecture, LiteDepthwiseNet, for HSI classification. Based on 3-D depthwise convolution, LiteDepthwiseNet can decompose standard convolution into depthwise convolution and pointwise convolution, which can achieve high classification performance with minimal parameters. Moreover, we remove the ReLU layer and batch normalization layer in the original 3-D depthwise convolution, which is likely to improve the overfitting phenomenon of the model on small-sized data sets. In addition, focal loss is used as the loss function to improve the model's attention on difficult samples and unbalanced data, and its training performance is significantly better than that of cross-entropy loss or balanced cross-entropy loss. Experiment results on five benchmark hyperspectral data sets show that LiteDepthwiseNet achieves state-of-the-art performance with a very small number of parameters and low computational cost.

Due to the detailed spectral information through hundreds of narrow spectral bands provided by hyperspectral image (HSI) data, it can be employed to accurately classify diverse materials of interest, which is one of the core applications of hyperspectral remote sensing technology. In recent years, with the rapid development of deep learning, convolutional neural networks (CNNs) have been successfully applied in many fields, including HSI classification. However, the random gradient descent-based parameter updating scheme is too general and leading to the inefficiency of CNN models. Moreover, the high dimensionality and limited training samples of HSI data also exacerbate the overfitting problem. To tackle these issues, in this article, a novel deep network with multilayer and multibranch architecture, named 3-D Gabor CNN (3DG-CNN), is proposed for HSI classification. More precisely, since the predefined 3-D Gabor filters in multiple scales and orientations could well characterize the internal spatial-spectral structure of HSI data from various perspectives, the 3-D Gabor-modulated kernels (3-D GMKs) are employed to replace the random initialization kernels. Moreover, the specially designed multibranch architecture enables the network to better integrating the scalable property of 3-D Gabor filters; thus, the representative ability and robustness of the extracted features can be greatly improved. Alternatively, the number of network parameters is substantially reduced due to the incorporation of 3-D Gabor modulation, relieving the training complexity and also alleviating the training process from overfitting. Experimental results on four real HSI datasets (including two newly released ones in the literature) have demonstrated that the proposed 3DG-CNN model can achieve better performance than several widely used machine-learning-based and deep-learning-based approaches. For the sake of reproducibility, the codes of the proposed 3DG-CNN model are available at http://jiasen.tech/papers/.

With the development of hyperspectral imaging technology, hyperspectral images (HSIs) have become important when analyzing the class of ground objects. In recent years, benefiting from the massive labeled data, deep learning has achieved a series of breakthroughs in many fields of research. However, labeling HSIs requires sufficient domain knowledge and is time-consuming and laborious. Thus, how to apply deep learning effectively to small labeled samples is an important topic of research in HSI classification. To solve this problem, we propose a semisupervised Siamese network that embeds Siamese network into a semisupervised learning scheme. It integrates an autoencoder module and a Siamese network to, respectively, investigate information in a large amount of unlabeled data and rectify it with a limited labeled sample set, which is called 3DAES. First, the autoencoder method is trained on the massive unlabeled data to learn the refinement representation, creating an unsupervised feature. Second, based on this unsupervised feature, limited labeled samples are used to train a Siamese network to rectify the unsupervised feature to improve feature separability among various classes. Furthermore, by training the Siamese network, a random sampling scheme is used to accelerate training and avoid imbalance among various sample classes. Experiments on three benchmark HSI datasets consistently demonstrate the effectiveness and robustness of the proposed 3DAES approach with limited labeled samples. For study replication, the code developed for this study is available at https://github.com/ShuGuoJ/3DAES.git.

Background A sterilizing cure of HIV-1 infection has been reported in 2 persons living with HIV-1 who underwent allogeneic hematopoietic stem cell transplantations from donors who were homozygous for the CCR5Δ32 gene polymorphism. However, this has been considered elusive during natural infection. Objective To evaluate persistent HIV-1 reservoir cells in an elite controller with undetectable HIV-1 viremia for more than 8 years in the absence of antiretroviral therapy. Design Detailed investigation of virologic and immunologic characteristics. Setting Tertiary care centers in Buenos Aires, Argentina, and Boston, Massachusetts. Patient A patient with HIV-1 infection and durable drug-free suppression of HIV-1 replication. Measurements Analysis of genome-intact and replication-competent HIV-1 using near-full-length individual proviral sequencing and viral outgrowth assays, respectively; analysis of HIV-1 plasma RNA by ultrasensitive HIV-1 viral load testing. Results No genome-intact HIV-1 proviruses were detected in analysis of a total of 1.188 billion peripheral blood mononuclear cells and 503 million mononuclear cells from placental tissues. Seven defective proviruses, some of them derived from clonally expanded cells, were detected. A viral outgrowth assay failed to retrieve replication-competent HIV-1 from 150 million resting CD4+ T cells. No HIV-1 RNA was detected in 4.5 mL of plasma. Limitations Absence of evidence for intact HIV-1 proviruses in large numbers of cells is not evidence of absence of intact HIV-1 proviruses. A sterilizing cure of HIV-1 can never be empirically proved. Conclusion Genome-intact and replication-competent HIV-1 were not detected in an elite controller despite analysis of massive numbers of cells from blood and tissues, suggesting that this patient may have naturally achieved a sterilizing cure of HIV-1 infection. These observations raise the possibility that a sterilizing cure may be an extremely rare but possible outcome of HIV-1 infection. Primary funding source National Institutes of Health and Bill & Melinda Gates Foundation.

The nonoverlapped spectrum range between low spatial resolution (LR) hyperspectral (HS) and high spatial resolution (HR) multispectral (MS) images has been a fundamental but challenging problem for MS/HS fusion. The spectrum of HS data is generally 400-2500 nm, and the spectrum of MS data is generally 400-900 nm; how to obtain the high-fidelity HR HS fused image within the whole spectrum of 400-2500 nm? In this article, we proposed a band divide-and-conquer framework (BDCF) to solve the problem, by comprehensively considering spectral fidelity, spatial enhancement, and computational efficiency. First, the spectral bands of HS were divided into overlapped and nonoverlapped bands according to the spectral response between HS and MS. Then, a novel improved component substitution (CS)-based method by combing neural network was proposed to fuse the overlapped bands of LR HS. Then, a mapping-based method with the neural network was presented to construct the complicated nonlinear relationship between overlapped and nonoverlapped bands of the original LR HS data. The trained network was mapped to the fused overlapped HR HS bands to estimate the nonoverlapped HR HS bands. Experimental results on two simulated data sets and two realistic data sets of Gaofen (GF)-5 LR HS, GF-1 MS, and Sentinel-2A MS show that the proposed BDCF has superior performance in both high spectral fidelity and sharp spatial details, and it obtained competitive fusion behaviors compared with other state-of-the-art methods. Moreover, BDCF has relatively higher computational efficiency than optimal solution-based methods and deep learning-based fusion methods.

This article proposes a multiscale spectral features graph fusion (MSFGF) method for selecting proper hyperspectral bands. The MSFGF regards that the selected bands should reflect diagnostic spectral information of ground objects at different scales, and it explores band selection from the aspect of multiple spatial scales. First, it adopts the multiscale low-rank decomposition (MSLRD) model to find multiscale spectral features of different ground objects. The model considers divergent spatial structures or spatial correlations of ground objects at different scales, and factorizes the hyperspectral data cube into a series of low-rank block-wise data cubes, where the blocks take spatial structures of different ground objects at increasing scales. Second, the MSFGF presents the multiscale sparse spectral clustering (MSSC) model to fuse the separate connected graphs of multiscale spectral features into a consensus graph. The consensus graph combines the complementary information of multiscale spectral features and helps to reveal the intrinsic clustering structure of all spectral bands. Finally, the MSFGF utilizes spectral clustering to find clusters from the consensus graph and selects representative bands. Experimental results on three widely used hyperspectral data prove the superiority of MSFGF in selecting bands, where it outperforms other seven state-of-the-art methods in classification with an acceptable computational cost.

Deep neural networks have achieved promising performance for hyperspectral image (HSI) classification. However, due to the limitation of the available labeled samples, the traditional deeper and wider neural networks usually cause the overfitting problem and lose the detailed information. To solve this problem, a brain-like structure, namely spatial attention-driven recurrent feedback convolutional neural network (SARFNN), is proposed by utilizing the recurrent feedback and attention mechanism structures, from which two deep models are further developed for HSI classification. First, a 2-D SARFNN (SARF2DNN) model is developed to learn the spatial features from HSI data. After that, to better exploit the 3-D characteristic, the 3-D version is extended from SARF2DNN, thus constructing an SARF3DNN model to extract joint spatial-spectral features. Moreover, with the help of the idea of brain-likeness, the recurrent feedback module is designed to recover information loss caused by deeper structure and the dimension reduction operation. The experimental results conducted on two HSI data sets show that our SARFNN architecture can achieve more competitive performance than other state-of-the-art algorithms.

Pansharpening aims to sharpen a low spatial resolution (LR) multispectral (MS) image using a high spatial resolution (HR) panchromatic (PAN) image to obtain the HR MS image. Though large numbers of pansharpening methods have been proposed, and many advanced methods have shown high quantitative results, few of them are widely used in real applications. This may be attributed to their instability for different images with different ground surface features, or the complexity to be implemented and the time-consuming process for some state-of-the-art methods. In this letter, we proposed a simple, adjustable, and robust fusion (SARF) method. In the proposed method, a spatial-spectral coenhanced strategy was proposed, and several details of the proposed fusion model were specifically designed for the ``simple, adjustable, robust'' features. It was tested and verified by four-band and eight-band MS images based on reduced resolution (RR) and full resolution (FR) experiments. The experimental results demonstrated the promising spatial visuality of the proposed method, and the spectral fidelity was more robust than most of component substitution (CS)-based and multiresolution analysis (MRA)-based methods.

This letter presents a multiscale low-rank decomposition (MSLRD) method to extract multiscale spatial structures from hyperspectral images. The MSLRD assumes that ground objects have divergent characteristics in changing spatial scales. It decomposes each band image into a series of block-wise matrices, where these low-rank blocks take detailed spatial structures at multiple scales. It formulates the low-rank matrix decomposition problem into minimizing the ranks of all block matrices and adopts the alternative direction of the multiplier method to optimize it. Experiments on Indian Pines and Pavia University data sets show that the MSLRD can greatly improve the classification performance of regular classification on spectral features (i.e., all bands) and perform better than five state-of-the-art spatial feature extraction methods.

The monitoring of coastal wetlands is of great importance to the protection of marine and terrestrial ecosystems. However, due to the complex environment, severe vegetation mixture, and difficulty of access, it is impossible to accurately classify coastal wetlands and identify their species with traditional classifiers. Despite the integration of multisource remote sensing data for performance enhancement, there are still challenges with acquiring and exploiting the complementary merits from multisource data. In this article, the depthwise feature interaction network (DFINet) is proposed for wetland classification. A depthwise cross attention module is designed to extract self-correlation and cross correlation from multisource feature pairs. In this way, meaningful complementary information is emphasized for classification. DFINet is optimized by coordinating consistency loss, discrimination loss, and classification loss. Accordingly, DFINet reaches the standard solution-space under the regularity of loss functions, while the spatial consistency and feature discrimination are preserved. Comprehensive experimental results on two hyperspectral and multispectral wetland datasets demonstrate that the proposed DFINet outperforms other competitive methods in terms of overall accuracy.

Nonnegative matrix factorization (NMF) is a widely used hyperspectral unmixing model which decomposes a known hyperspectral data matrix into two unknown matrices, i.e., endmember matrix and abundance matrix. Due to the use of least-squares loss, the NMF model is usually sensitive to noise or outliers. To improve its robustness, we introduce a general robust loss function to replace the traditional least-squares loss and propose a general loss-based NMF (GLNMF) model for hyperspectral unmixing in this letter. The general loss function is a superset of many common robust loss functions and is suitable for handling different types of noise. Experimental results on simulated and real hyperspectral data sets demonstrate that our GLNMF model is more accurate and robust than existing NMF methods.

Gastric cancer is one of the most common cancers, which ranks third among the leading causes of cancer death. Biopsy of gastric mucosa is a standard procedure in gastric cancer screening test. However, manual pathological inspection is labor-intensive and time-consuming. Besides, it is challenging for an automated algorithm to locate the small lesion regions in the gigapixel whole-slide image and make the decision correctly. To tackle these issues, we collected large-scale whole-slide image dataset with detailed lesion region annotation and designed a whole-slide image analyzing framework consisting of 3 networks which could not only determine the screening result but also present the suspicious areas to the pathologist for reference. Experiments demonstrated that our proposed framework achieves sensitivity of 97.05 % and specificity of 92.72 % in screening task and Dice coefficient of 0.8331 in segmentation task. Furthermore, we tested our best model in real-world scenario on 10 , 315 whole-slide images collected from 4 medical centers.

Recently, multiscale spatial features have been widely utilized to improve the hyperspectral image (HSI) classification performance. However, fixed-size neighborhood involving the contextual information probably leads to misclassifications, especially for the boundary pixels. Additionally, it has been demonstrated that deep neural network (DNN) is practical to extract representative features for the classification tasks. Nevertheless, under the condition of high dimensionality versus small sample sizes, DNN tends to be over-fitting and it is generally time-consuming due to the deep-level feature learning process. To alleviate the aforementioned issues, we propose a multiscale context-aware ensemble deep kernel extreme learning machine (MSC-EDKELM) for efficient HSI classification. First, the scene of the HSI data set is over-segmented in multiscale via using the adaptive superpixel segmentation technique. Second, superpixel pattern (SP) and attentional neighboring superpixel pattern (ANSP) are generated by leveraging the superpixel maps, which can automatically comprise local and global contextual information, respectively. Afterward, an ensemble deep kernel extreme learning machine (EDKELM) is presented to investigate the deep-level characteristics in the SP and ANSP. Finally, the category of each pixel is accurately determined by the decision fusion and weighted output layer fusion strategy. Experimental results on four real-world HSI data sets demonstrate that the proposed frameworks outperform some classic and state-of-the-art methods with high computational efficiency, which can be employed to serve real-time applications.

Background With the increase of the number of smokers, tobacco exposure among pregnant women is becoming more and more common. Pregnant women exposed to first-hand smoke and second-hand smoke are susceptible to physiological and psychological health issues has been proved in previous studies. Nevertheless, there are no enough studies focus on the impact of third-hand smoke during pregnancy. This study aimed to assess and compare health-related quality of life for pregnant women with exposure to first-hand smoke, second-hand smoke, third-hand smoke and non-exposure to tobacco in mainland China. Methods National-based cross-sectional study is based on a questionnaire survey which collects information including demographics, smoking behaviors and self-evaluation. All questionnaires were delivered and collected from August to September 2019. EuroQol group’s visual analog scale and EuroQoL Five-dimension Questionnaire were used to collect data in mainland China. Results Totally, 15,682 pregnant women were included in this study, among which non-exposure to smoke were 7564 (48.2%), exposed to first-hand smoke, second-hand smoke and third-hand smoke were 89 (0.6%), 2349 (15.0%), and 5680 (36.2%) respectively. Pregnant women without tobacco exposure had the highest EuroQol group’s visual analog scale score (mean value = 85.4[SD = 14.0]), while those with first-hand smoke had the lowest score (mean value = 77.4[SD = 22.2]). Among all five dimensions of EuroQoL Five-dimension Questionnaire, there were significant differences of EQ-index among groups with different tobacco exposure in usual activity and anxiety or depression dimensions (p Conclusions Third-hand smoke exposure had close relationship with low health-related quality of life in pregnant women. Moreover, second-hand smoke exposure significantly led more problems on mental dimension of pregnant women.

In this work, we present a phenomenon-oriented comparative analysis of the two dominant approaches in English Resource Semantic (ERS) parsing: classic, knowledge-intensive and neural, data-intensive models. To reflect state-of-the-art neural NLP technologies, a factorization-based parser is introduced that can produce Elementary Dependency Structures much more accurately than previous data-driven parsers. We conduct a suite of tests for different linguistic phenomena to analyze the grammatical competence of different parsers, where we show that, despite comparable performance overall, knowledge- and data-intensive models produce different types of errors, in a way that can be explained by their theoretical properties. This analysis is beneficial to in-depth evaluation of several representative parsing techniques and leads to new directions for parser development.

Pansharpening aims to sharpen a lowspatial-resolution (LR) multispectral (MS) image using a high-spatial-resolution (HR) panchromatic (Pan) image to obtain the HR MS image. It has been a fundamental and active research topic in remote sensing, and pansharpening methods have been developed for nearly 40 years. While the performance evaluation of pansharpening methods is still based on a small number of individual images, datadriven pansharpening approaches are attracting increasing attention. However, few publicly available benchmark data sets for pansharpening are available, especially large-scale ones. This has been a serious limitation for the future development of pansharpening methods.

The presence of mixed pixels in the hyperspectral data makes unmixing to be a key step for many applications. Unsupervised unmixing needs to estimate the number of endmembers, their spectral signatures, and their abundances at each pixel. Since both endmember and abundance matrices are unknown, unsupervised unmixing can be considered as a blind source separation problem and can be solved by nonnegative matrix factorization (NMF). However, most of the existing NMF unmixing methods use a least-squares objective function that is sensitive to the noise and outliers. To deal with different types of noises in hyperspectral data, such as the noise in different bands (band noise), the noise in different pixels (pixel noise), and the noise in different elements of hyperspectral data matrix (element noise), we propose three self-paced learning based NMF (SpNMF) unmixing models in this article. The SpNMF models replace the least-squares loss in the standard NMF model with weighted least-squares losses and adopt a self-paced learning (SPL) strategy to learn the weights adaptively. In each iteration of SPL, atoms (bands or pixels or elements) with weight zero are considered as complex atoms and are excluded, while atoms with nonzero weights are considered as easy atoms and are included in the current unmixing model. By gradually enlarging the size of the current model set, SpNMF can select atoms from easy to complex. Usually, noisy or outlying atoms are complex atoms that are excluded from the unmixing model. Thus, SpNMF models are robust to noise and outliers. Experimental results on the simulated and two real hyperspectral data sets demonstrate that our proposed SpNMF methods are more accurate and robust than the existing NMF methods, especially in the case of heavy noise.

Deep learning (DL) algorithms have been demonstrated to have great potential in hyperspectral image (HSI) classification. However, most DL methods mainly focus on improving classification performance, neglecting the computational cost. In order to broaden the application scenarios of DL-based HSI classification methods, it is necessary to develop lightweight and fast models to fit the deployment needs of computational-resource-limited platforms. Taking into account the efficiency and accuracy, this article designs a lightweight network architecture based on spectral and channel-wise attention modules, namely LiteSCANet, for HSI classification. It contains a residual double-branch structure, which makes the model effectively extract spectral-spatial features and achieve good performance with fast inference speed and low computational consumption (i.e., floating-point operations, number of parameters, and graphics processing unit memory usage). The experiment results on four benchmark datasets show that our proposed model achieves an excellent trade-off between efficiency and accuracy compared with the other six existing networks.

This article presents a label similarity probability filter (LSPF) to make hyperspectral image postclassification. The LSPF is inspired by the first law of geography and proposes a class label probability function to quantify the probability of both centered and its neighboring pixels belonging to the same class. It first classifies the hyperspectral data using the regular support vector machine classifier. Then, it binarizes the posterior classification result to obtain the binary label maps of each class. After that, it traverses all spatial windows centered by each pixel and calculates the cumulative probability of all pixels in each class. Finally, the cumulative probabilities are used to make reclassification to obtain the refined classification map. The experiments on Indian Pines, Pavia University, and ZY1-02D Yellow River Estuary data show that LSPF greatly improves the classification accuracy of spectral signatures and outperforms other state-of-the-art spectral–spatial methods.

Recent studies in cross-domain classification have shown that discriminant information of both source and target domains is very important. In this article, we propose a new domain adaptation (DA) method for hyperspectral image (HSI) classification, called graph embedding and distribution alignment (GEDA). GEDA uses the graph embedding method and a pseudo-label learning method to learn interclass and intraclass divergence matrices of source and target domains, which preserves the local discriminant information of both domains. Meanwhile, spatial and spectral features of HSI are used, and distribution alignment and subspace alignment are performed to minimize the spectral differences between domains. We perform DA tasks on Yancheng, Botswana, University of Pavia, and Center of Pavia, Shanghai and Hangzhou data sets. Experimental results show that the classification performance of the proposed GEDA is better than that of existing DA methods.

Owing to the complicated and heterogeneous distribution characteristics of wetland features, the existing hyperspectral technology is difficult to investigate the inner-pixel subtle changes. In this article, we present a subpixel change detection method based on collaborative coupled unmixing (SCDUM) for monitoring coastal wetlands. A novel multitemporal and spatial scale collaborative endmember extraction method based on joint spatial and spectral information is proposed. In the proposed method, the multitemporal hyperspectral images are first jointly clustered and segmented based on multifeature fusion of spectral features, texture features, and shape features. Then, a different spatial scale nonnegative matrix factorization based on original and downsampled multitemporal hyperspectral images is proposed to accurately extract the pure endmembers of each segmented images. Finally, the global abundance of the multitemporal image is effectively estimated for change detection. In addition, in order to verify the accuracy of the change detection results without reference, an accuracy verification strategy by using high spatial resolution Sentinel-2A image as auxiliary data is implemented. The Yellow River Estuary coastal wetland was selected as the research area, and the Gaofen-5 and ZY-1 02D hyperspectral images were used as the research data. In particular, the proposed method not only provides the overall change information, but also obtains the component of change direction and intensity of each kind of endmember, and the experimental results show that the SCDUM gives more accurate detection results, with closer to the endmember spectral curves of real objects, compared with other state-of-the-art methods.

Background: The aim of this study was to estimate the attributable risk for all-cause mortality in hypertensive adults living in Beijing, China. Methods: We conducted a prospective cohort study on the basis of the disease risk prediction model, which included 3006 hypertensive patients aged 50 and over who participated in the annual health examination from thirty-eight community health centers were randomly selected from all 53 community health centers in Dongcheng district of Beijing in China. This cohort study was conducted from January 1, 2013 to June 31, 2018 in these community health centers. Data included age, gender, education level, BMI, smoking and drinking status, renal function, diabetes mellitus (DM), coronary heart disease, levels of blood pressure, use of medications, and blood lipid levels. Results: the follow-up time was 4.90 ± 0.51 years. There were significant survival differences by gender, renal function (eGFR > 90 vs. 60–90 vs. 12 years), coronary heart disease (CHD) (CHD vs. NO CHD). In the multivariate Cox proportional hazard analysis, the prognostic factors of all-cause mortality in hypertensive patients were male [HR 1.662, 95% CI 1.110–2.489, p = 0.014], educational level

This article presents a fast and latent low-rank subspace clustering (FLLRSC) method to select hyperspectral bands. The FLLRSC assumes that all the bands are sampled from a union of latent low-rank independent subspaces and formulates the self-representation property of all bands into a latent low-rank representation (LLRR) model. The assumption ensures sufficient sampling bands in representing low-rank subspaces of all bands and improves robustness to noise. The FLLRSC first implements the Hadamard random projections to reduce spatial dimensionality and lower the computational cost. It then adopts the inexact augmented Lagrange multiplier algorithm to optimize the LLRR program and estimates sparse coefficients of all the projected bands. After that, it employs a correntropy metric to measure the similarity between pairwise bands and constructs an affinity matrix based on sparse representation. The correntropy metric could better describe the nonlinear characteristics of hyperspectral bands and enhance the block-diagonal structure of the similarity matrix for correctly clustering all subspaces. The FLLRSC conducts spectral clustering on the connected graph denoted by the affinity matrix. The bands that are closest to their separate cluster centroids form the final band subset. Experimental results on three widely used hyperspectral data sets show that the FLLRSC performs better than the classical low-rank representation methods with higher classification accuracy at a low computational cost.

Highly sensitive visual detection of proteins in space is often complicated, since bioassay payloads must achieve sensitive detection and environmental adaptability. Microfluidic chips enabling high degrees of integration and automation are a very promising approach for highly sensitive visualized protein detection in space. Here, we present a chip-based, highly integrated payload technology for the visual detection of proteins onboard the Tianzhou-1 cargo spacecraft. This device simply and tightly integrated multifunctional bioassay chips, complex fluid control, and execution of multiple programs to provide highly sensitive protein detection. Based on the delicate design of the microfluidic network, superparamagnetic immunoassay particles were successfully used to achieve immunoaffinity enrichment and visual detection in chips. This method demonstrated good performance compared to traditional visual test strips, showing limits of detection in the picomolar range for interleukin (IL)-6. Results measured in orbit indicated rapid visual detection of IL-6 in cell culture media, confirming the validity of the proposed approach for protein detection in space.

Reversible image data hiding (RIDH) is a special class of data hiding techniques, where the host image can be perfectly reconstructed upon data extraction. Due to this reversibility property, RIDH has been widely adopted in many critical scenarios. However, almost all the existing methods focus on improving the capacity-distortion performance; and the hiding ability is ambiguously referred as the perceptual unawareness of a human observer. In this brief, we show that the prevalent RIDH framework, prediction error expansion histogram shifting (PEE-HS), would leave quite obvious traces after embedding, suffering the risks to expose the data hiding action. To address this issue, several countermeasures are proposed to conceal the embedding traces while retaining the conventional reversibility feature. The experimental results demonstrate the effectiveness of our proposed method. We believe this brief could shed light on the security aspect of RIDH.

OBJECTIVE: To investigate whether salidroside (Sal) protected the rat heart from exhaustive exercise-induced injury by inducing nuclear factor erythroid 2-related factor 2 (Nrf2) expression. METHODS: Forty-eight male Sprague-Dawley rats were divided into four groups (n = 12 rats per group): the control, the exhaustive swimming (ES) group, the low-dose Sal plus acute exhaustive swimming (SLE) group, and the high-dose Sal plus acute exhaustive swimming (SHE) group. In the SLE and SHE groups, 15 and 30 mg/kg Sal were administered, respectively, once a day. The rats in the control and ES groups were administered the same amount of physiological saline, respectively, once a day. On the 14th day, the rats in the ES, SLE and SHE groups underwent exhaustive swimming training once. Then cardiac function parameters and electrocardiograms were recorded. Biomarkers of myocardial injury in the serum and oxidative stress factors in the myocardial tissue were evaluated using ELISA tests. The levels of Nrf2, nuclear Nrf2 and Kelch-like ECH-associated protein 1 (Keap1) messenger RNA and proteins were assessed in the myocardium using q-PCR and Western blotting, respectively. RESULTS: Compared to the control group, the ES group showed remarkable increases in serum brain natriuretic peptide (BNP), cardiac troponin I (cTnI) and reactive oxygen species levels, but significant decreases in catalase and glutathione levels (p < 0.05). Compared to the ES group, the Sal treatment decreased serum BNP and cTnI levels and alleviated the changes in levels of oxidative stress-related factors. After treatment with Sal, nuclear and intracellular levels of Nrf2 protein were increased in the myocardial cells, while the level of Keap1 protein was decreased (p < 0.05). CONCLUSION: Sal protected the heart from exhaustive exerciseinduced injury, and it may improve cardiac function and cardiac bioelectricity in exhausted rats by inducing Nrf2 expression.

This letter presents a correntropy-based sparse spectral clustering (CSSC) method to select proper bands of a hyperspectral image. The CSSC first constructs an affinity matrix with the correntropy measure which considers the nonlinear characteristics of hyperspectral bands and can suppress effects from noise or outliers in measuring band similarity. The CSSC imposes the sparsity and block diagonal constraint on spectral clustering, which can further improve band clustering performance. Bands are finally selected from each cluster on the connected graph. Experimental results on two widely used hyperspectral images show that the CSSC behaves better than spectral clustering and other several state-of-the-art methods in band selection.

This letter proposes a lateral-slice sparse tensor robust principal component analysis (LSSTRPCA) method to remove gross errors or outliers from hyperspectral images so as to promote the performance of subsequent classification. The LSSTRPCA assumes that a three-order hyperspectral tensor has a low-rank structure, and gross errors or outliers are sparsely scattered in a 2-D space (i.e., lateral-slice) of the tensor. It formulates a low-rank and sparse tensor decomposition problem into a convex problem and then implements the inexact augmented Lagrange multiplier method to solve it. The experiments on two hyperspectral data sets show that the LSSTRPCA can successfully remove outliers or gross errors and achieve higher accuracies than both the original robust principal component analysis (RPCA) and tensor robust principal component analysis (TRPCA).

With the development of various deep learning algorithms, the importance and potential of AI + medical treatment are increasingly prominent. Electrocardiogram (ECG) as a common auxiliary diagnostic index of heart diseases, has been widely applied in the pre-screening and physical examination of heart diseases due to its low price and non-invasive characteristics. Currently, the multi-lead ECG equipments have been used in the clinic, and some of them have the automatic analysis and diagnosis functions. However, the automatic analysis is not accurate enough for the discrimination of abnormal events of ECG, which needs to be further checked by doctors. We therefore develop a deep-learning-based approach for multi-label classification of ECG named Multi-ECGNet, which can effectively identify patients with multiple heart diseases at the same time. The experimental results show that the performance of our methods can get a high score of 0.863 (micro-F1-score) in classifying 55 kinds of arrythmias, which is beyond the level of ordinary human experts.

This article proposes a novel unsupervised domain adaptation (DA) method called ideal regularized discriminative multiple kernel subspace alignment (IRDMKSA) for hyperspectral image (HSI) classification. The proposed IRDMKSA method includes three main steps: ideal regularization, discriminative multiple kernel learning, and subspace alignment. The ideal regularization strategy exploits label information of source domain to refine the standard source and target kernels and also to build a connection between them. The discriminative multiple kernel learning can learn a composite kernel to describe the nonlinearity of HSI samples by fusing complementary information among different single kernels. Finally, the subspace alignment is used to diminish the difference between source and target composite kernels. The proposed IRDMKSA method exploits both the sample similarity and label similarity and makes the resulting kernel more appropriate for DA tasks. Experimental results on four DA tasks show that the performance of IRDMKSA is better than some classical unsupervised DA methods for the HSI classification.

Background Transcriptome data generates massive amounts of information that can be used for characterization and prognosis of patient outcomes for many diseases. The goal of our research is to predict the survival time of lung adenocarcinoma patients and improve the accuracy of classifying the long-survival cohort and short-survival cohort. Methods We filtered prognostic features related with survival time of lung adenocarcinoma patients by the method of Relief and predicted whether survival time of the patient is >3 years or not—using eight machine learning algorithms (Support Vector Machines, Random Forests, Logistic Regression, Naïve Bayes, Linear Regression, Support Vector Regression (kernel Poly), Support Vector Regression (kernel Linear), and Ridge Regression). Then the best-performed algorithm was chosen to build a predictive model of survival time of lung adenocarcinoma patients. Further, another dataset was used to verify the stability and suitability of this model. We explored the underlying mechanisms of RNA expression changes with the corresponding DNA mutations and DNA methylation patterns in the 22 selected genetic features. Results The best machine learning algorithm was Naïve Bayes (accuracy=75%, AUC =0.81) using the top 22 genetic features, and this algorithm had the stable and great performance on another dataset as well. The coupled mutation number of the long-survival group (>6 years) was less than the short-survival group (

Clinical cases of chromosome 7 long-arm end deletion are rare. Generally, 7q terminal deletion syndrome results in complex clinical phenotypes, such as microcephaly, growth and development retardation, holoprosencephaly, and sacral hypoplasia. Herein, we report the genetic and clinical features of a fetus with multiple malformations observed by prenatal ultrasound. The results showed that there was a large fragment deletion of approximately 27.7 Mb in 7q32.3-qter. The induced fetus showed facial abnormalities of cleft lip and palate, and some organ structural abnormalities (such as diaphragmatic hernia and polycystic renal dysplasia) were observed by autopsy and pathology. To provide more reliable information for disease diagnosis and genetic counseling, we reviewed and analyzed the reported cases of isolated 7q terminal syndrome.

To improve the reliability of current signal monitoring tool wear status, a method based on Mahalanobis distance to identify tool wear status is proposed. First, the obtained current signal is analysed in time domain, frequency domain and wavelet domain, and several features that have good correlation with tool wear status are selected to form the feature vector. The feature vector of the current signal in normal tool wear status is taken as the reference vector. Then calculate the Mahalanobis distance value of the feature vectors of the current signal of the tool with moderate wear and severe wear, so that two corresponding thresholds T(1) and T(2) can be obtained. The feature vector of an unknown wear status is calculated using the Mahalanobis distance and then compared with the two thresholds obtained previously. When the calculated value is between threshold T(1) and T(2), the tool is judged to be in moderate wear status. When its Mahalanobis distance value is greater than the threshold T(2), it is judged that the tool has been seriously worn. Finally, multiple unknown wear status is identified. It is believed that the recognition method based on Mahalanobis distance can accurately determine the tool wear status.

Background With the rapid development of the socioeconomic status, the mortality of several cancers has been changed in China during the past 30 years. We aimed to estimate the trends of mortality and years of life lost (YLLs) of various cancers in urban and rural areas of China from 1990 to 2017. Methods The mortality data were collected from Chinese yearbooks and the age structure of population from the Chinese sixth population census were used as reference to calculate age‐standardized mortality rates (ASMRs) and YLLs rates. Joinpoint regression analysis was implemented to calculate the annual percent change (APC) of mortality rates and YLL rates for cancers. YLLs owing to premature death were calculated as age‐specific cancer deaths multiplied by the reference life expectancy at birth of 80 years for male and 82.5 years for female. Results The ASMRs of all cancers showed significant decreasing trends for urban residents from 1990 to 2017, such downward trend without significance was also observed among rural residents. Interestingly, ASMRs of lung cancer and breast cancer have raised continuously in rural areas since 1990. The age‐standardized YYL rates for urban and rural residents decreased with 1.02% and 0.85% per year, respectively. YLLs in rural areas were higher than those in urban areas, whereas YLLs of urban outstripped those of rural finally with the increasing in YLLs of urban areas (216.71% for men and 207.87% for women). Conclusion The ASMRs and YLL rates of all cancers have declined in urban and rural areas from 1990 to 2017. YLLs increased in urban areas and remained higher level in rural areas after 2014 year. Preventive measures should be strengthened to against cancer, especially for lung cancer., Mortality rate and YLL rate of cancer were analyzed by joinpoint regression. The ASMRs of cancer in both urban and rural areas in China declined from 1990 to 2017. YLLs increased in urban areas and remained higher levels in rural areas from 1990 to 2017.

Motivated by the superior performance of region covariance descriptor, we use covariance matrices as new features to replace the original spectral pixel features, and employ a Log-Euclidean metric to characterize the geodesic distance between symmetric positive definite (SPD) covariance matrices. Based on the covariance features and Log-Euclidean metric, we propose a Log-Euclidean kernel-based joint sparse representation (LogEKJSR) model for the classification of hyperspectral images (HSIs). In the implementation of LogEKJSR, we first reduce the dimensionality of HSI by employing extended multiattribute profile (EMAP) transformations, and then extract the region covariance matrix features associated with each pixel on the EMAPs. The EMAP can model homogeneity and texture structure of HSI by aggregating multiple morphological attribute profiles, the covariance matrix feature contains both local spectral correlation and spatial structure information, and the Log-Euclidean kernel is a matrix-to-matrix similarity metric rather than vector-to-vector similarity metric. Finally, a LogEKJSR model is obtained by replacing the traditional kernel in the KJSR with the Log-Euclidean kernel. Experimental results on three benchmark hyperspectral data sets demonstrate that our proposed LogEKJSR is more effective than existing JSR, KJSR, and support vector machine methods.

Diffuse cutaneous systemic sclerosis may occur in women of childbearing age. Pregnancies in this population are associated with a markedly increased risk of adverse obstetric and maternal outcomes even before the onset of symptoms related to sclerosis. We report a case involving the management and outcome of pregnancy in a 30-year-old woman with diffuse cutaneous systemic sclerosis. The course of her pregnancy was good and was assisted by a group consultation including obstetricians and rheumatologists. Vaginal delivery was the patient’s preferred choice because she had irregular skin tightness in her lower abdominal skin. She underwent induction of labor and combined spinal-epidural analgesia, and successfully delivered. Importantly, these pregnancies need to be planned, where possible, to allow the opportunity to counsel women and their partners in advance and to decrease any risks. These pregnancies should be considered high risk, and they require close antenatal monitoring and good supervision from an expert multidisciplinary team experienced in high-risk pregnancies. The management of delivery for patients with cutaneous systemic sclerosis is challenging, and vaginal delivery with labor analgesia is an alternative option to cesarean section.

Osteonecrosis of the femoral head (ONFH) is a leading cause of mobility impairment which may lead to a total hip replacement. Recent studies have found tendon derived stem cells (TDSCs) might be an ideal cell source for musculoskeletal tissue regeneration. And our previous study has shown Sox11 could promote osteogenesis of bone marrow-derived MSCs. However, the effect of TDSCs or Sox11 over-expressing TDSCs (TDSCs-Sox11) on bone regeneration in ONFH has not been investigated. In the present study, TDSCs were infected with AAV carrying Sox11 or empty vector. We showed that Sox11 could promote the proliferation and osteogenic differentiation of TDSCs, as well as angiogenesis in vitro. The western blot analysis showed that Sox11 could activate the PI3K/Akt signaling pathway to promote osteogenesis of TDSCs. Finally, using a rabbit model of hormone-induced ONFH, our result demonstrated that local administration of TDSCs or TDSCs overexpressing Sox11 could accelerate bone regeneration in necrotic femoral heads, and TDSCs overexpressing Sox11 showed better effects. TDSCs over-expressing Sox11 might be a promising cell source for stem cell therapy to promote bone regeneration, such as ONFH, fracture, bone defect, and so on.

Background : Hepatocellular carcinoma (HCC) had high prevalence and poor prognosis, and hepatitis B virus (HBV) infection is a major risk factor. The aim of this study is to analyze the role of long intergenic noncoding RNA 01232 (LINC01232) in the prognosis and progression of HCC, and explore the relationship between LINC01232 and HBV infection. Methods : LINC01232 expression and its prognostic value were firstly analyzed using TCGA database. Quantitative real-time PCR was used to evaluate the expression of LINC01232 in HCC patients and cell lines. Kaplan-Meier curves were used to analyze the relationship between LINC01232 expression and HCC overall survival prognosis. Function-loss in vitro experiments were performed to demonstrate the role of LINC01232 in HCC progression. A luciferase reporter assay and Pearson correlation were used to confirm the relationship between LINC01232 and microRNA (miR)-708-5p in HCC. Results : The expression of LINC01232 was upregulated in HCC tissues and cell lines, and high LINC01232 was associated with worse overall survival in HCC. LINC01232 reduction inhibited HCC cells proliferation, migration and invasion. LINC01232 expression was significantly correlated with HBV infection and liver cirrhosis, and showed potential to distinguish HBV-infected HCC patients. miR-708-5p, as a HBV-related miRNA, was a potential target of LINC01232, and was negatively correlated with LINC01232 in HCC. Conclusion : Our findings found that highly expressed LINC01232 may be a biomarker to indicate survival prognosis in HCC patients, especially in HBV-infected cases. In addition, LINC01232 plays as an oncogene in HCC progression, and its function may exert by sponging miR-708-5p.

Today's open-domain conversational agents increase the informativeness of generated responses by leveraging external knowledge. Most of the existing approaches work only for scenarios with a massive amount of monolingual knowledge sources. For languages with limited availability of knowledge sources, it is not effective to use knowledge in the same language to generate informative responses. To address this problem, we propose the task of cross-lingual knowledge grounded conversation (CKGC), where we leverage large-scale knowledge sources in another language to generate informative responses. Two main challenges come with the task of cross-lingual knowledge grounded conversation: (1) knowledge selection and response generation in a cross-lingual setting; and (2) the lack of a test dataset for evaluation. To tackle the first challenge, we propose the curriculum self-knowledge distillation (CSKD) scheme, which utilizes a large-scale dialogue corpus in an auxiliary language to improve cross-lingual knowledge selection and knowledge expression in the target language via knowledge distillation. To tackle the second challenge, we collect a cross-lingual knowledge grounded conversation test dataset to facilitate relevant research in the future. Extensive experiments on the newly created dataset verify the effectiveness of our proposed curriculum self-knowledge distillation method for cross-lingual knowledge grounded conversation. In addition, we find that our proposed unsupervised method significantly outperforms the state-of-the-art baselines in cross-lingual knowledge selection.

We propose and demonstrate a high performance four-wavelength erbium-doped fiber laser (EDFL), enabled by a figure-8 compound-ring-cavity (F8-CRC) filter for single-longitudinal-mode (SLM) selection and a polarization-managed four-channel filter (PM-FCF) for defining four lasing wavelengths. We introduce a novel methodology utilizing signal-flow graph combined with Mason's rule to analyze a CRC filter in general and apply it to obtain the important design parameters for the F8-CRC filter used in this paper. By combining the functions of the F8-CRC filter and the PM-FCF filter assisted by the enhanced polarization hole-burning and polarization dependent loss, we achieve the EDFL with fifteen lasing states, including four single-, six dual-, four tri- and one quad-wavelength lasing operations. In particular, all the four single-wavelength operations are in stable SLM oscillation, typically with a linewidth of =3 MHz and an output power fluctuation of, 21 pages, 14 figures

Background Pregnant women experience physical, physiological, and mental changes. Health-related quality of life (HRQoL) is a relevant indicator of psychological and physical behaviours, changing over the course of pregnancy. This study aims to assess HRQoL of pregnant women during different stages of pregnancy. Methods This cross-sectional study was performed using the The EuroQoL Group’s five-dimension five-level questionnaire (EQ-5D-5L) to assess the HRQoL of pregnant women, and demographic data were collected. This study was conducted in a regional university hospital in Guangzhou, China. Results A total of 908 pregnant women were included in this study. Pregnant women in the early 2nd trimester had the highest HRQoL. The HRQoL of pregnant women rose from the 1st trimester to the early 2nd trimester, and dropped to the bottom at the late 3rd trimester due to some physical and mental changes. Reports of pain/discomfort problem were the most common (46.0%) while self-care were the least concern. More than 10% of pregnant women in the 1st trimester had health-related problems in at least one dimension of whole five dimensions. In the whole sample, the EuroQoL Group’s visual analog scale (EQ-VAS) was 87.86 ± 9.16. Across the gestational stages, the HRQoL remained stable during the pregnancy but the highest value was observed in the 1st trimester (89.65 ± 10.13) while the lowest was in the late 3rd trimester (87.28 ± 9.13). Conclusions During pregnancy, HRQoL were associated with gestational trimesters in a certain degree. HRQoL was the highest in the early 2nd trimester and then decreased to the lowest in the late 3rd trimester due to a series of physical and psychological changes. Therefore, obstetric doctors and medical institutions should give more attention and care to pregnant women in the late 3rd trimester.

A hyperspectral imaging sensor collects detailed spectral responses from ground objects using hundreds of narrow bands; this technology is used in many real-world applications. Band selection aims to select a small subset of hyperspectral bands to remove spectral redundancy and reduce computational costs while preserving the significant spectral information of ground objects. In this article, we review current hyperspectral band selection methods, which can be classified into six main categories: ranking based, searching based, clustering based, sparsity based, embedding-learning based, and hybrid-scheme based. With two widely used hyperspectral data sets, we illustrate the classification performances of several popular band selection methods. The challenges and research directions of hyperspectral band selection are also discussed.

Domain adaptation, which aims at learning an accurate classifier for a new domain (target domain) using labeled information from an old domain (source domain), has shown promising value in remote sensing fields yet still been a challenging problem. In this letter, we focus on knowledge transfer between hyperspectral remotely sensed images in the context of land-cover classification under unsupervised setting where labeled samples are available only for the source image. Specifically, a discriminative transfer joint matching (DTJM) method is proposed, which matches source and target features in the kernel principal component analysis space by minimizing the empirical maximum mean discrepancy, performs instance reweighting by imposing an $\ell _{2,1}$ -norm on the embedding matrix, and preserves the local manifold structure of data from different domains and meanwhile maximizes the dependence between the embedding and labels. The proposed approach is compared with some state-of-the-art feature extraction techniques with and without using label information of source data. Experimental results on two benchmark hypersepctral data sets show the effectiveness of the proposed DTJM.

A new framework is proposed for the fusion of hyperspectral and light detection and ranging (LiDAR) data based on the extinction profiles (EPs), local binary pattern (LBP), and kernel collaborative representation classification. Specifically, EP and LBP features are extracted from both sources. Then, the derived features of each source are classified by collaborative representation-based classifier with Tikhonov regularization (KCRT). Reconstruction residuals are fused to produce the final label assignment. Experimental results demonstrate that the proposed method outperforms the existing methods in hyperspectral and LiDAR data fusion.

With the advancement of social media and mobile technology, any smartphone user can easily become a seller on social media and e-commerce platforms, such as Instagram and Carousell in Hong Kong or Taobao in China. A seller shows images of their products and annotates their images with suitable tags that can be searched easily by others. Those images could be taken by the seller, or the seller could use images shared by other sellers. Among sellers, some sell counterfeit goods, and these sellers may use disguising tags and language, which make detecting them a difficult task. This article proposes a framework to detect counterfeit sellers by using deep learning to discover connections among sellers from their shared images. Based on 473K shared images from Taobao, Instagram, and Carousell, it is proven that the proposed framework can detect counterfeit sellers. The framework is 30% better than approaches using object recognition in detecting counterfeit sellers. To the best of our knowledge, this is the first work to detect online counterfeit sellers from their shared images.

Skutterudite crystals are of interest for their thermoelectric properties and in particular for the flexibility the structure offers for tuning such properties. One strategy to enhance the thermoel...

MXenes (Mn+1Xn, e.g., Ti3C2) are the largest 2D material family developed in recent years. They exhibit significant potential in the energy sciences, particularly for energy storage. In this review, we summarize the progress of the computational work regarding the theoretical design of new MXene structures and predictions for energy applications including their fundamental, energy storage, and catalytic properties. We also outline how high-throughput computation, big data, and machine-learning techniques can help broaden the MXene family. Finally, we present some of the major remaining challenges and future research directions needed to mature this novel materials family.