Search

Abstract Knowledge Distillation has shown promising results for classifying skin lesions in dermatology images. Traditional knowledge distillation typically involves the student model passively mimicking the teacher model's knowledge. We propose utilizing AdaBoost to enable the student to actively mine the teacher's learning representation for skin lesion classification. This paradigm allows the student to determine the “granularity” in mining the teacher's knowledge. As the student's learning process progresses, it can become challenging to pinpoint specific learning difficulties, especially with potential interference from the teacher. To address this issue, we introduce a variational difficulty mining strategy to reduce the impact of such interference. This strategy involves the distillation module capturing more nuanced classification difficulties by extracting information from the node's $$l$$ l th hops. By maximizing the mutual information between the teacher and student, we effectively filter out noise interference from these nuanced difficulties. Our proposed framework, Variational AdaBoost Knowledge Distillation (VAdaKD), allows the student to actively mine and leverage the teacher's knowledge for improved skin lesion classification. Our proposed method performs satisfactorily on three benchmark datasets: the Dermnet dataset, ISIC 2019 dataset, and HAM10000 dataset, respectively. Specifically, our method shows an improvement of 2–3% over the baseline on the Dermnet dataset and outperforms the best results of the other compared methods by 1%. Experimental results and visualization performance indicate that our proposed method effectively captures the learning difficulties and achieves better visualized t-distributed stochastic neighbor embedding classification results. Our code is available at https://github.com/25brilliant/VAdaKD .

The success of vision Transformers (ViTs) relies heavily on the self-attention mechanism, which requires support from appropriate patch tokenization. However, hyperspectral image (HSI) often suffer from significant noise distortions and spectral uncertainty, which result in unstable attention patterns and overfitting due to equivocal tokenization. In this paper, we propose a neighborhood contrastive tokenization task (NeiCoT) to learn compact, semantically meaningful, and context-sensitive tokens for efficient Transformer encoding. Specifically, we employ a predictor on patch embedding to maximize the mutual information between local individuals and their global average anchor. This encourages neighboring tokens’ relevance and active participation in feature learning. Next, we revise a token-level contrastive loss to align predictions with local individuals and distinguish them from other samples in a mini-batch to enhance tokens rich in contextual semantics. Furthermore, we apply a Gaussian weighting to the tokens’ contrastive loss to balance the neighborhood contribution. Finally, we propose a sequence-specific MAE framework with NeiCoT to achieve HSI representation, and additionally validate NeiCoT on a supervised Transformer backbone. The results demonstrate that NeiCoT consistently enhances the robustness and generalization of the Transformer, achieving accurate object recognition and boundary localization even with limited training samples. Our code will be available at https://github.com/zoegnov07/NeiCoT.

Following publication, concerns were raised regarding the peer-review process related to the publication of this article [...]

BackgroundMultiple regimens have been widely used in the eradication treatment of Helicobacter pylori infection in children. However, there is a lack of comparison and evaluation of their effectiveness in different regions of the world.MethodsRandomized controlled trials were retrieved. Review Manager 5.4, Stata SE 15 and R 4.0.4 statistical software were used to analyze date. The ranking probability is assessed according to the surfaces under cumulative ranking (SUCRA).Results163 studies were eligible for this study, involving 336 arms and 18,257 children, and 10 different interventions. The results showed that the eradication rates of sequential therapy with probiotics (SP), bismuth-containing quadruple (Quadruple) therapy, concomitant therapy and PCN therapy were at least 90%. Cumulative ranking showed that SP therapy had the best eradication effect (SUCRA 92.7%) whereas Bismuth-containing triple therapy (B) had the worst (SUCRA 3.5%). Subgroup analysis suggested that SP therapy ranked first in China and other regions, and the ranking of Triple therapy with probiotics therapy (TP) was equally stable (SUCRA 72.0% vs 76.4% respectively). The security of the SP and TP therapy had great advantages.ConclusionsAs for the eradication treatment of Helicobacter pylori infection in children, SP therapy ranks highest. SP and TP therapies are most safe.

Background: Necroptosis and inflammation are closely related to the pathogenesis of respiratory syncytial virus (RSV). Acteoside (AC), a natural phenylpropanoid glycoside from Kuding Tea, has significant anti-RSV effect. However, the roles of AC on RSV-induced lung necroptosis and inflammation are yet to be elucidated.Methods: The effects of AC were investigated in BALB/c mice and A549 cells. Lung histopathology was observed through H&E staining. The viral titer was assessed via plaque assay. The RSV-F expression was determined by RT-qPCR and immunohistochemistry assay. The levels of cytokines were detected by ELISA and RT-qPCR. The necroptosis rate and mitochondrial membrane potential were evaluated via flow cytometry. The expressions of HMGB1/NF-κB and RIP1/RIP3/MLKL/PGAM5/DRP1 were detected by western blot. Additionally, untargeted metabolomics was conducted to investigate the metabolic profiles and related metabolic pathways via Gas Chromatography-Mass Spectrometry.Results: The results showed that compared with the RSV-infected group, AC treatment significantly attenuated lung pathological damage, virus replication, and cytokines levels. AC also alleviated RSV-induced necroptosis and mitochondrial dysfunction in vitro and in vivo. Moreover, AC treatment down-regulated the expression of HMGB1, p-Iκbα/Iκbα, p-p65/p65, RIP1, RIP3, MLKL, PGAM5, and DRP1. Furthermore, metabolomic analyses suggested that the perturbations in major metabolites of AC therapy were related to variations in amino acid and energy metabolism.Conclusion: Our findings validated the beneficial effects of AC in suppressing necroptosis and regulating metabolism, suggesting AC may be a new drug candidate for RSV infection.

Convolutional neural networks (CNNs) have improved the accuracy of hyperspectral image (HSI) classification significantly. However, CNN models usually generate a large number of feature maps, which lead to high redundancy and cannot guarantee to effectively extract discriminative features for well characterizing the complex structures of HSIs. In this article, two novel mixed link networks (MLNets) are proposed to enhance the representational ability of CNNs for HSI classification. Specifically, the proposed mixed link architectures integrate the feature reusage property of the residual network and the capability of effective new feature exploration of the densely convolutional network, extracting more discriminative features from HSIs. Compared with the dual path architecture, the proposed mixed link architectures can further improve the information flow throughout the network. Experimental results on three hyperspectral benchmark datasets demonstrate that our MLNets achieve competitive results compared with other state-of-the-art HSI classification approaches.

Objective: Major depressive disorder (MDD) is one of the most common psychiatric disorders, the diagnosis and treatment of MDD are major clinical issues. However, there is a lack of effective biomarkers and drugs diagnosis and therapeutics of MDD. In the present study, bioinformatics analysis combined with an experimental verification strategy was used to identify biomarkers and paeoniflorin targets for MDD diagnosis and treatment.Methods: Based on network pharmacology, we obtained potential targets and pathways of paeoniflorin as an antidepressant through multiple databases. We then constructed a protein-protein interaction network and performed enrichment analyses. According to the results, we performed in vivo and in vitro experimental validation.Results: The results showed that paeoniflorin may exert an antidepressant effect by regulating cell inflammation, synaptic function, NF-κB signaling pathway, and intestinal inflammation.Conclusion: NPM1, HSPA8, HSPA5, HNRNPU, and TNF are the targets of paeoniflorin treatment. In addition, we demonstrated that paeoniflorin inhibits inflammatory cytokine production via the p38MAPK/NF-κB pathway and has neuroprotective effects on the synaptic structure. Our findings provide valuable evidence for the diagnosis and treatment of MDD.

In recent years, semantic segmentation with pixel-level classification has become one of the types of research focus in the field of polarimetric synthetic aperture radar (PolSAR) image interpretation. Fully convolutional network (FCN) can achieve end-to-end semantic segmentation, which provides a basic framework for subsequent improved networks. As a classic FCN-based network, U-Net has been applied to semantic segmentation of remote sensing images. Although good segmentation results have been obtained, scalar neurons have made it difficult for the network to obtain multiple properties of entities in the image. The vector neurons used in the capsule network can effectively solve this problem. In this paper, we propose a complex-valued (CV) U-Net with a CV capsule network embedded for semantic segmentation of a PolSAR image. The structure of CV U-Net is lightweight to match the small PolSAR data, and the embedded CV capsule network is designed to extract more abundant features of the PolSAR image than the CV U-Net. Furthermore, CV dynamic routing is proposed to realize the connection between capsules in two adjacent layers. Experiments on two airborne datasets and one Gaofen-3 dataset show that the proposed network is capable of distinguishing different types of land covers with a similar scattering mechanism and extracting complex boundaries between two adjacent land covers. The network achieves better segmentation performance than other state-of-art networks, especially when the training set size is small.

Supported Pt-based catalysts have been identified as highly selective catalysts for CO oxidation, but their potential for applications has been hampered by the high cost and scarcity of Pt metals as well as aggregation problems at relatively high temperatures. In this work, nanorod structured (TiO2−Pt)/CeO2 catalysts with the addition of 0.3 at% Pt and different atomic ratios of Ti were prepared through a combined dealloying and calcination method. XRD, XPS, SEM, TEM, and STEM measurements were used to confirm the phase composition, surface morphology, and structure of synthesized samples. After calcination treatment, Pt nanoparticles were semi-inlayed on the surface of the CeO2 nanorod, and TiO2 was highly dispersed into the catalyst system, resulting in the formation of (TiO2−Pt)/CeO2 with high specific surface area and large pore volume. The unique structure can provide more reaction path and active sites for catalytic CO oxidation, thus contributing to the generation of catalysts with high catalytic activity. The outstanding catalytic performance is ascribed to the stable structure and proper TiO2 doping as well as the combined effect of Pt, TiO2, and CeO2. The research results are of importance for further development of high catalytic performance nanoporous catalytic materials.

Gametogenesis is an essential step for malaria parasite transmission and is activated in mosquito by signals including temperature drop, pH change, and mosquito-derived xanthurenic acid (XA). Recently, a membrane protein gametogenesis essential protein 1 (GEP1) was found to be responsible for sensing these signals and interacting with a giant guanylate cyclase α (GCα) to activate the cGMP-PKG-Ca2+ signaling pathway for malaria parasite gametogenesis. However, the molecular mechanisms for this process remain unclear. In this study, we used AlphaFold2 to predict the structure of GEP1 and found that it consists of a conserved N-terminal helical domain and a transmembrane domain that adopts a structure similar to that of cationic amino acid transporters. Molecular docking results showed that XA binds to GEP1 via a pocket similar to the ligand binding sites of known amino acid transporters. In addition, truncations of this N-terminal sequence significantly enhanced the expression, solubility, and stability of GEP1. In addition, we found that GEP1 interacts with GCα via its C-terminal region, which is interrupted by mutations of a few conserved residues. These findings provide further insights into the molecular mechanism for the XA recognition by GEP1 and the activation of the gametogenesis of malaria parasites through GEP1-GCα interaction.

In China, food has become safer over the past five years, especially commonly consumed foods. Food supervision sampling has played an important role in improving food safety. However, consumer acceptance of the results of food safety supervision have not kept pace. Communicating actual food safety risks to consumers and improving the public trust in food safety supervision sampling inspection has become an important issue. This study focused on food safety surveillance sampling of commonly consumed foods. In total, 4408 adult consumers were surveyed between August and October 2021. Structural equation modeling was performed for data analysis. This study found significant differences along gender lines and across different cities and levels of education with respect to evaluating competence trust and care trust on food supervision sampling inspection. This study identified the public’s competence trust, care trust, and perception of food safety as factors that significantly affect one’s attitude toward supervision sampling inspection. Care trust showed a more pronounced effect on trust enhancement than competence trust. The present study also provides some practical measures for food safety supervisors to improve public trust in the national food inspection. Specifically, the sampling process should be open and transparent.

To learn discriminative features, hyperspectral image (HSI), containing 3-D cube data, is a preferable means of capturing multi-head self-attention from both spatial and spectral domains if the burden in model optimization and computation is low. In this paper, we design a dual multi-head contextual self-attention (DMuCA) network for HSI classification with the fewest possible parameters and lower computation costs. To effectively capture rich contextual dependencies from both domains, we decouple the spatial and spectral contextual attention into two sub-blocks, SaMCA and SeMCA, where depth-wise convolution is employed to contextualize the input keys in the pure dimension. Thereafter, multi-head local attentions are implemented as group processing when the keys are alternately concatenated with the queries. In particular, in the SeMCA block, we group the spatial pixels by evenly sampling and create multi-head channel attention on each sampling set, to reduce the number of the training parameters and avoid the storage increase. In addition, the static contextual keys are fused with the dynamic attentional features in each block to strengthen the capacity of the model in data representation. Finally, the decoupled sub-blocks are weighted and summed together for 3-D attention perception of HSI. The DMuCA module is then plugged into a ResNet to perform HSI classification. Extensive experiments demonstrate that our proposed DMuCA achieves excellent results over several state-of-the-art attention mechanisms with the same backbone.

Hyperspectral images (HSIs), acquired as a 3D data set, contain spectral and spatial information that is important for ground–object recognition. A 3D convolutional neural network (3DCNN) could therefore be more suitable than a 2D one for extracting multiscale neighborhood information in the spectral and spatial domains simultaneously, if it is not restrained by mass parameters and computation cost. In this paper, we propose a novel lightweight multilevel feature fusion network (LMFN) that can achieve satisfactory HSI classification with fewer parameters and a lower computational burden. The LMFN decouples spectral–spatial feature extraction into two modules: point-wise 3D convolution to learn correlations between adjacent bands with no spatial perception, and depth-wise convolution to obtain local texture features while the spectral receptive field remains unchanged. Then, a target-guided fusion mechanism (TFM) is introduced to achieve multilevel spectral–spatial feature fusion between the two modules. More specifically, multiscale spectral features are endowed with spatial long-range dependency, which is quantified by central target pixel-guided similarity measurement. Subsequently, the results obtained from shallow to deep layers are added, respectively, to the spatial modules, in an orderly manner. The TFM block can enhance adjacent spectral correction and focus on pixels that actively boost the target classification accuracy, while performing multiscale feature fusion. Experimental results across three benchmark HSI data sets indicate that our proposed LMFN has competitive advantages, in terms of both classification accuracy and lightweight deep network architecture engineering. More importantly, compared to state-of-the-art methods, the LMFN presents better robustness and generalization.

Convolutional neural networks (CNNs) are the go-to model for hyperspectral image (HSI) classification because of the excellent locally contextual modeling ability that is beneficial to spatial and spectral feature extraction. However, CNNs with a limited receptive field pose challenges for modeling long-range dependencies. To solve this issue, we introduce a novel classification framework which regards the input HSI as a sequence data and is constructed exclusively with multilayer perceptrons (MLPs). Specifically, we propose a spectral-spatial MLP (SS-MLP) architecture, which uses matrix transposition and MLPs to achieve both spectral and spatial perception in global receptive field, capturing long-range dependencies and extracting more discriminative spectral-spatial features. Four benchmark HSI datasets are used to evaluate the classification performance of the proposed SS-MLP. Experimental results show that our pure MLP-based architecture outperforms other state-of-the-art convolution-based models in terms of both classification performance and computational time. When comparing with the SSSERN model, the average accuracy improvement of our approach is as high as 3.03%. We believe that our impressive experimental results will foster additional research on simple yet effective MLP-based architecture for HSI classification.

Convolutional neural networks (CNNs) have achieved great results in hyperspectral image (HSI) classification in recent years. However, convolution kernels are reused among different spatial locations, known as spatial-agnostic or weight-sharing kernels. Furthermore, the preference of spatial compactness in convolution (typically, 3×3 kernel size) constrains the receptive field and the ability to capture long-range spatial interactions. To mitigate the above two issues, in this article, we combine a novel operation called involution with residual learning and develop a new deep residual involution network (DRIN) for HSI classification. The proposed DRIN could model long-range spatial interactions well by adopting enlarged involution kernels and realize feature learning in a fairly lightweight manner. Moreover, the vast and dynamic involution kernels are distinct over different spatial positions, which could prioritize the informative visual patterns in the spatial domain according to the spectral information of the target pixel. The proposed DRIN achieves better classification results when compared with both traditional machine learning-based and convolution-based methods on four HSI datasets. Especially in comparison with the convolutional baseline model, i.e., deep residual network (DRN), our involution-powered DRIN model increases the overall classification accuracy by 0.5%, 1.3%, 0.4%, and 2.3% on the University of Pavia, the University of Houston, the Salinas Valley, and the recently released HyRANK HSI benchmark datasets, respectively, demonstrating the potential of involution for HSI classification.

The convolutional neural network (CNN) can automatically extract hierarchical feature representations from raw data and has recently achieved great success in the classification of hyperspectral images (HSIs). However, most CNN based methods used in HSI classification neglect adequately utilizing the strong complementary yet correlated information from each convolutional layer and only employ the last convolutional layer features for classification. In this paper, we propose a novel fully dense multiscale fusion network (FDMFN) that takes full advantage of the hierarchical features from all the convolutional layers for HSI classification. In the proposed network, shortcut connections are introduced between any two layers in a feed-forward manner, enabling features learned by each layer to be accessed by all subsequent layers. This fully dense connectivity pattern achieves comprehensive feature reuse and enforces discriminative feature learning. In addition, various spectral-spatial features with multiple scales from all convolutional layers are fused to extract more discriminative features for HSI classification. Experimental results on three widely used hyperspectral scenes demonstrate that the proposed FDMFN can achieve better classification performance in comparison with several state-of-the-art approaches.

Filter banks transferred from a pre-trained deep convolutional network exhibit significant performance in heightening the inter-class separability for hyperspectral image feature extraction, but weakening the intra-class consistency simultaneously. In this paper, we propose a new superpixel-based relational auto-encoder for cohesive spectral−spatial feature learning. Firstly, multiscale local spatial information and global semantic features of hyperspectral images are extracted by filter banks transferred from the pre-trained VGG-16. Meanwhile, we utilize superpixel segmentation to construct the low-dimensional manifold embedded in the spectral domain. Then, representational consistency constraint among each superpixel is added in the objective function of sparse auto-encoder, which iteratively assist and supervisedly learn hidden representation of deep spatial feature with greater cohesiveness. Superpixel-based local consistency constraint in this work not only reduces the computational complexity, but builds the neighborhood relationships adaptively. The final feature extraction is accomplished by collaborative encoder of spectral−spatial feature and weighting fusion of multiscale features. A large number of experimental results demonstrate that our proposed method achieves expected results in discriminant feature extraction and has certain advantages over some existing methods, especially on extremely limited sample conditions.

Convolutional neural networks (CNNs) have recently shown outstanding capability for hyperspectral image (HSI) classification. In this work, a novel CNN model is proposed, which is wider than other existing deep learning-based HSI classification models. Based on the fact that very deep residual networks (ResNets) behave like ensembles of relatively shallow networks, our proposed network, called multipath ResNet (MPRN), employs multiple residual functions in the residual blocks to make the network wider, rather than deeper. The proposed network consists of shorter-medium paths for efficient gradient flow and replaces the stacking of multiple residual blocks in ResNet with fewer residual blocks but more parallel residual functions in each of it. Experimental results on three real hyperspectral data sets demonstrate the superiority of the proposed method over several state-of-the-art classification methods.

Pt/CeO2 catalysts with nanoporous structures were prepared by the facile dealloying of melt-spun Al92−XCe8PtX (X = 0.1; 0.3 and 0.5) ribbons followed by calcination. The phase compositions and structural parameters of the catalysts were characterized by X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM) and high-resolution transmission electron microscopy (HRTEM). The specific surface area and pore size distribution were characterized by N2 adsorption−desorption tests. The catalytic properties were evaluated by a three-way catalyst (TWC) measurement system. The results revealed that the dealloyed samples exhibited a nanorod framework structure. The Pt nanoparticles that formed in situ were supported and highly dispersed on the CeO2 nanorod surface and had sizes in the range of 2−5 nm. For the catalyst prepared from the melt-spun Al91.7Ce8Pt0.3 ribbons, the 50% CO conversion temperature (T50) was 91 °C, and total CO could be converted when the temperature was increased to 113 °C. An X-ray photoelectron spectroscopy (XPS) test showed that the Pt0.3/CeO2 sample had a slightly richer oxygen vacancy; and a H2 temperature-programmed reduction (H2-TPR) test demonstrated its superior adsorption ability for reduction gas and high content of active oxygen species. The experiments indicated that the catalytic performance could be retained without any attenuation after 130 h when water and CO2 were present in the reaction gas. The favorable catalytic activities were attributed to the high specific areas and small pore and Pt particle sizes as well as the strong interactions between the CeO2 and Pt nanoparticles. The Pt nanoparticles were embedded in the surface of the CeO2 nanorods, inhibiting growth. Therefore, the catalytic stability and water resistance were excellent.