Search Results (1,415)

The semantic segmentation of satellite and UAV remote sensing imagery is pivotal for address exploration, change detection, quantitative analysis and urban planning. Recent advancements have seen an influx of segmentation networks utilizing convolutional neural networks and transformers. However, the intricate geographical features and varied land cover boundary interferences in remote sensing imagery still challenge conventional segmentation networks’ spatial representation and long-range dependency capabilities. This paper introduces a novel U-Net-like network for UAV image segmentation. We developed a link aggregation Mamba at the critical skip connection stage of UNetFormer. This approach maps and aggregates multi-scale features from different stages into a unified linear dimension through four Mamba branches containing state-space models (SSMs), ultimately decoupling and fusing these features to restore the contextual relationships in the mask. Moreover, the Mix-Mamba module is incorporated, leveraging a parallel self-attention mechanism with SSMs to merge the advantages of a global receptive field and reduce modeling complexity. This module facilitates nonlinear modeling across different channels and spaces through multipath activation, catering to international and local long-range dependencies. Evaluations on public remote sensing datasets like LovaDA, UAVid and Vaihingen underscore the state-of-the-art performance of our approach. Full article

Depth information is crucial for perceiving three-dimensional scenes. However, depth maps captured directly by depth sensors are often incomplete and noisy, our objective in the depth-completion task is to generate dense and accurate depth maps from sparse depth inputs by fusing guidance information from corresponding color images obtained from camera sensors. To address these challenges, we introduce transformer models, which have shown great promise in the field of vision, into the task of image-guided depth completion. By leveraging the self-attention mechanism, we propose a novel network architecture that effectively meets these requirements of high accuracy and resolution in depth data. To be more specific, we design a dual-branch model with a transformer-based encoder that serializes image features into tokens step by step and extracts multi-scale pyramid features suitable for pixel-wise dense prediction tasks. Additionally, we incorporate a dual-attention fusion module to enhance the fusion between the two branches. This module combines convolution-based spatial and channel-attention mechanisms, which are adept at capturing local information, with cross-attention mechanisms that excel at capturing long-distance relationships. Our model achieves state-of-the-art performance on both the NYUv2 depth and SUN-RGBD depth datasets. Additionally, our ablation studies confirm the effectiveness of the designed modules. Full article

Recently, Vision Transformers (ViTs) have been actively applied to fine-grained visual recognition (FGVR). ViT can effectively model the interdependencies between patch-divided object regions through an inherent self-attention mechanism. In addition, patch selection is used with ViT to remove redundant patch information and highlight the most discriminative object patches. However, existing ViT-based FGVR models are limited to single-scale processing, and their fixed receptive fields hinder representational richness and exacerbate vulnerability to scale variability. Therefore, we propose MultiScale Patch Selection (MSPS) to improve the multiscale capabilities of existing ViT-based models. Specifically, MSPS selects salient patches of different scales at different stages of a MultiScale Vision Transformer (MS-ViT). In addition, we introduce Class Token Transfer (CTT) and MultiScale Cross-Attention (MSCA) to model cross-scale interactions between selected multiscale patches and fully reflect them in model decisions. Compared with previous Single-Scale Patch Selection (SSPS), our proposed MSPS encourages richer object representations based on feature hierarchy and consistently improves performance from small-sized to large-sized objects. As a result, we propose M2Former, which outperforms CNN-/ViT-based models on several widely used FGVR benchmarks. Full article

Falls are the second leading cause of unintentional injury deaths worldwide. While numerous wearable fall detection devices incorporating AI models have been developed, none of them are used successfully in a fall detection application running on commodity-based smartwatches in real time. The system misses some falls, and generates an annoying amount of False Positives for practical use. We have investigated and experimented with an LSTM model for fall detection on a smartwatch. Even though the LSTM model has high accuracy during offline testing, the good performance of offline LSTM models cannot be translated to the equivalence of real-time performance. Transformers, on the other hand, can learn long-sequence data and patterns intrinsic to the data due to their self-attention mechanism. This paper compares three variants of LSTM and two variants of Transformer models for learning fall patterns. We trained all models using fall and activity data from three datasets, and the real-time testing of the model was performed using the SmartFall App. Our findings showed that in the offline training, the CNN-LSTM model was better than the Transformer model for all the datasets. However, the Transformer is a preferable choice for deployment in real-time fall detection applications. Full article

Synthetic Aperture Radar (SAR) imaging is essential for monitoring geomorphic changes, urban transformations, and natural disasters. However, the inherent complexities of SAR, particularly pronounced speckle noise, often lead to numerous false detections. To address these challenges, we propose the Multidirectional Attention Fusion Network (MDAF-Net), an advanced framework that significantly enhances image quality and detection accuracy. Firstly, we introduce the Multidirectional Filter (MF), which employs side-window filtering techniques and eight directional filters. This approach supports multidirectional image processing, effectively suppressing speckle noise and precisely preserving edge details. By utilizing deep neural network components, such as average pooling, the MF dynamically adapts to different noise patterns and textures, thereby enhancing image clarity and contrast. Building on this innovation, MDAF-Net integrates multidirectional feature learning with a multiscale self-attention mechanism. This design utilizes local edge information for robust noise suppression and combines global and local contextual data, enhancing the model’s contextual understanding and adaptability across various scenarios. Rigorous testing on six SAR datasets demonstrated that MDAF-Net achieves superior detection accuracy compared with other methods. On average, the Kappa coefficient improved by approximately 1.14%, substantially reducing errors and enhancing change detection precision. Full article

This paper investigates the feasibility of downscaling within high-dimensional Lorenz models through the use of machine learning (ML) techniques. This study integrates atmospheric sciences, nonlinear dynamics, and machine learning, focusing on using large-scale atmospheric data to predict small-scale phenomena through ML-based empirical models. The high-dimensional generalized Lorenz model (GLM) was utilized to generate chaotic data across multiple scales, which was subsequently used to train three types of machine learning models: a linear regression model, a feedforward neural network (FFNN)-based model, and a transformer-based model. The linear regression model uses large-scale variables to predict small-scale variables, serving as a foundational approach. The FFNN and transformer-based models add complexity, incorporating multiple hidden layers and self-attention mechanisms, respectively, to enhance prediction accuracy. All three models demonstrated robust performance, with correlation coefficients between the predicted and actual small-scale variables exceeding 0.9. Notably, the transformer-based model, which yielded better results than the others, exhibited strong performance in both control and parallel runs, where sensitive dependence on initial conditions (SDIC) occurs during the validation period. This study highlights several key findings and areas for future research: (1) a set of large-scale variables, analogous to multivariate analysis, which retain memory of their connections to smaller scales, can be effectively leveraged by trained empirical models to estimate irregular, chaotic small-scale variables; (2) modern machine learning techniques, such as FFNN and transformer models, are effective in capturing these downscaling processes; and (3) future research could explore both downscaling and upscaling processes within a triple-scale system (e.g., large-scale tropical waves, medium-scale hurricanes, and small-scale convection processes) to enhance the prediction of multiscale weather and climate systems. Full article

In recent years, graph similarity algorithms have been extensively developed based on neural networks. However, with an increase in the node count in graphs, these models either suffer from a reduced representation ability or face a significant increase in the computational cost. To address this issue, a graph similarity algorithm based on graph partitioning and attention mechanisms was proposed in this study. Our method first divided each input graph into the subgraphs to directly extract the local structural features. The residual graph convolution and multihead self-attention mechanisms were employed to generate node embeddings for each subgraph, extract the feature information from the nodes, and regenerate the subgraph embeddings using varying attention weights. Initially, rough cosine similarity calculations were performed on all subgraph pairs from the two sets of subgraphs, with highly similar pairs selected for precise similarity computation. These results were then integrated into the similarity score for the input graph. The experimental results indicated that the proposed learning algorithm outperformed the traditional algorithms and similar computing models in terms of graph similarity computation performance. Full article

Recently, deep unfolding network methods have significantly progressed in hyperspectral snapshot compressive imaging. Many approaches directly employ Transformer models to boost the feature representation capabilities of algorithms. However, they often fall short of leveraging the full potential of self-attention mechanisms. Additionally, current methods lack adequate consideration of both intra-stage and inter-stage feature fusion, which hampers their overall performance. To tackle these challenges, we introduce a novel approach that hybridizes the sparse Transformer and wavelet fusion-based deep unfolding network for hyperspectral image (HSI) reconstruction. Our method includes the development of a spatial sparse Transformer and a spectral sparse Transformer, designed to capture spatial and spectral attention of HSI data, respectively, thus enhancing the Transformer’s feature representation capabilities. Furthermore, we incorporate wavelet-based methods for both intra-stage and inter-stage feature fusion, which significantly boosts the algorithm’s reconstruction performance. Extensive experiments across various datasets confirm the superiority of our proposed approach. Full article

To address the challenges of change detection tasks, including the scarcity and dispersion of labeled samples, the difficulty in efficiently extracting features from unstructured image objects, and the underutilization of high-order correlation information, we propose a novel architecture based on hypergraph convolutional neural networks. By characterizing superpixel vertices and their high-order correlations, the method implicitly expands the number of labels while assigning adaptive weight parameters to adjacent objects. It not only describes changes in vertex features but also uncovers local and consistent changes within hyperedges. Specifically, a vertex aggregation mechanism based on superpixel segmentation is established, which segments the difference map into superpixels of diverse shapes and boundaries, and extracts their significant statistical features. Subsequently, a dynamic hypergraph structure is constructed, with each superpixel serving as a vertex. Based on the multi-head self-attention mechanism, the connection probability between vertices and hyperedges is calculated through learnable parameters, and the hyperedges are generated through threshold filtering. Moreover, a framework based on hypergraph convolutional neural networks is customized, which models the high-order correlations within the data through the learning optimization of the hypergraph, achieving change detection in remote sensing images. The experimental results demonstrate that the method obtains impressive qualitative and quantitative analysis results on the three remote sensing datasets, thereby verifying its effectiveness in enhancing the robustness and accuracy of change detection. Full article

In brain–computer interface motor imagery (BCI-MI) systems, convolutional neural networks (CNNs) have traditionally dominated as the deep learning method of choice, demonstrating significant advancements in state-of-the-art studies. Recently, Transformer models with attention mechanisms have emerged as a sophisticated technique, enhancing the capture of long-term dependencies and intricate feature relationships in BCI-MI. This research investigates the performance of EEG-TCNet and EEG-Conformer models, which are trained and validated using various hyperparameters and bandpass filters during preprocessing to assess improvements in model accuracy. Additionally, this study introduces EEG-TCNTransformer, a novel model that integrates the convolutional architecture of EEG-TCNet with a series of self-attention blocks employing a multi-head structure. EEG-TCNTransformer achieves an accuracy of 83.41% without the application of bandpass filtering. Full article

Electric motors play a crucial role in self-driving vehicles. Therefore, fault diagnosis in motors is important for ensuring the safety and reliability of vehicles. In order to improve fault detection performance, this paper proposes a motor fault diagnosis method based on vibration signals. Firstly, the vibration signals of each operating state of the motor at different frequencies are measured with vibration sensors. Secondly, the characteristic of Gram image coding is used to realize the coding of time domain information, and the one-dimensional vibration signals are transformed into grayscale diagrams to highlight their features. Finally, the lightweight neural network Xception is chosen as the main tool, and the attention mechanism Convolutional Block Attention Module (CBAM) is introduced into the model to enforce the importance of the characteristic information of the motor faults and realize their accurate identification. Xception is a type of convolutional neural network; its lightweight design maintains excellent performance while significantly reducing the model’s order of magnitude. Without affecting the computational complexity and accuracy of the network, the CBAM attention mechanism is added, and Gram’s corner field is combined with the improved lightweight neural network. The experimental results show that this model achieves a better recognition effect and faster iteration speed compared with the traditional Convolutional Neural Network (CNN), ResNet, and Xception networks. Full article

Based on motivational theory, this study investigated the effect of users’ knowledge seeking on users’ knowledge contribution in question-and-answer (Q&A) communities. We collected 4643 samples from the largest social Q&A platform in China (Zhihu) and applied a mediation effect test to the data. The results showed that knowledge seeking affects intrinsic motivations (altruism and self-efficacy) and extrinsic motivations (social support, group identity, and reputation), further affecting knowledge contribution. Our findings indicated that Q&A communities should be concerned with users’ intrinsic and extrinsic motivations to ensure balanced knowledge exchange on social Q&A platforms, ultimately fostering long-term stability and growth. Existing research has mainly focused on a single behavioral state, such as knowledge seeking or knowledge contribution, and has paid little attention to the connection between these two types of user information behaviors. This study aimed to fill this gap by revealing the mechanisms through which users’ knowledge seeking affects their knowledge contribution. Full article

The prediction of fuel consumption and Carbon Intensity Index (CII) of ships is crucial for optimizing decarbonization strategies in the maritime industry. This study proposes a ship fuel consumption prediction model based on the Long Short-Term Memory with Self-Attention Mechanism (SA-LSTM). The model is applied to a container ship of 2400 TEU to predict its hourly fuel consumption, hourly CII, and annual CII rating. Four different feature sets are selected from these data sources and are used as inputs for SA-LSTM and another ten models. The results demonstrate that the SA-LSTM model outperforms the other models in prediction accuracy. Specifically, the Mean Absolute Percentage Error (MAPE) for fuel consumption predictions using the SA-LSTM model is reduced by up to 20% compared to the XGBoost and by up to 12% compared to the LSTM model. Additionally, the SA-LSTM model achieves the highest accuracy in annual CII predictions. Full article

Currently, less attention is paid to zinc–carbon batteries, although they are still widely used and are among the major types of batteries collected and recycled. The recycling technologies currently in use do not allow the complete recovery of resources, are not self-sufficient and require additional financing. Therefore, this paper aims to study the possibility of complete recycling of waste zinc–carbon batteries and to suggest the practical use of the final products generated in the recycling process. The possibility of complex processing of spent zinc–carbon batteries using mechanical separation and processing of the battery’s components (steel case, zinc electrode, graphite electrode, polypropylene and paper insulators) is justified. The separation of spent electrolytes from other components of batteries with hydrochloric acid was studied. It was shown that the extraction of Zn²⁺ and NH⁴⁺ cations takes place following the addition of an equivalent amount of Na₃PO₄ solution and water-insoluble NH₄ZnPO₄ salt sedimentation. Waste agglomerate (mixture of MnO₂, MnO(OH), and graphite) was regenerated to its initial composition (MnO₂, graphite) at a temperature of 300–325 °C; manganese (III) hydroxide was oxidized to manganese (IV) dioxide. Thermal destruction of polypropylene and paper insulators with additional introduction of polyethylene into the primary mixture produced pyrolysis liquid, pyrocarbon and pyrolysis gas as products. The practical use of the products obtained and compliance with the environmental requirements of the suggested method of waste batteries recycling were shown. Full article

When dealing with non-IID data, federated learning confronts issues such as client drift and sluggish convergence. Therefore, we propose a Bidirectional Corrective Model-Contrastive Federated Adversarial Training (BCMCFAT) framework. On the client side, we design a category information correction module to correct biases caused by imbalanced local data by incorporating the local client’s data distribution information. Through local adversarial training, more robust local models are obtained. Secondly, we propose a model-based adaptive correction algorithm in the server that leverages a self-attention mechanism to handle each client’s data distribution information and introduces learnable aggregation tokens. Through the self-attention mechanism, model contrast learning is conducted on each client to obtain aggregation weights of corrected client models, thus addressing the issues of accuracy degradation and slow convergence caused by client drift. Our algorithm achieves the best natural accuracy on the CIFAR-10, CIFAR-100, and SVHN datasets and demonstrates excellent adversarial defense performance against FGSM, BIM, and PGD attacks. Full article