Cloud obscuration undermines the availability of optical images for continuous monitoring in earth observation. Fusing features from synthetic aperture radar (SAR) has been recognized as a feasible strategy to guide the reconstruction of ...

• Propose a W-shaped asymmetric hierarchical framework for multimodal cloud removal.
• Design an interactive attention transformer for cross-modality feature fusion.
• Integrate multisensor features in both spatial and frequency domains ...

Single-modality medical images generally do not contain enough information to reach an accurate and reliable diagnosis. For this reason, physicians commonly rely on multimodal medical images for comprehensive diagnostic assessments. This study ...

• A new deep framework for multi-modality medical image segmentation is introduced.
• The approach is based on Dempster–Shafer theory of evidence.
• Mass functions are transformed by contextual discounting with learnt coefficients.
• ...

The learnable PatchMatch formulations have recently made progress in Multi-View Stereo (MVS). However, their performance often degrades under complex wild scenes. In this study, we observe that the degradation is mainly caused by the noisy depth ...

• A DefLearn strategy is devised to enable scene-aware depth hypothesis sampling learning.
• A PPMNet is proposed with DefLearn to address noisy depth estimation problem in the wild.
• PPMNet achieves SOTA performance constantly in ...

Federated learning (FL) facilitates a privacy-preserving neural network training paradigm through collaboration between edge clients and a central server. One significant challenge is that the distributed data is not independently and identically ...

• Federated learning preserves privacy via client–server collaborative training.
• A major challenge is data heterogeneity, divided into intra-and inter-domain.
• Cross-client local clustered feature contrast promotes knowledge-sharing.

This review comprehensively investigates the current state and emerging trends of autonomous vehicle terrain detection and segmentation. By systematically reviewing literature from various databases, this study outlines the evolution of detection ...

• Explores terrain detection and segmentation technologies in autonomous vehicles.
• Evaluates current methods’ effectiveness and limitations across environments.
• Identifies challenges in advancing terrain detection and segmentation ...

We focus on exploring an unsupervised learning-based model which can take advantage of a single image and events to estimate dense and time-continuous optical flow.

We propose a multi-scale optical flow recurrent estimation ...

• We propose a multi-scale optical flow recurrent estimation network called MREIFlow.
• We propose a hierarchical unsupervised learning scheme.
• We conduct experiments with MREIFlow and existing methods on the MVSEC dataset.

For image fusion tasks, it is inefficient for the diffusion model to iterate multiple times on the original resolution image for feature mapping. To address this issue, this paper proposes an efficient latent feature-guided diffusion model for ...

• A novel latent diffusion model is proposed for general image fusion.
• A UNet-based compression strategy adapts to various fusion tasks without pretraining.
• The Transformer model boosts denoising performance as a new diffusion ...

The expansion of the Internet has led to the widespread proliferation of malicious URLs, becoming a primary vector for cyber threats. Detecting malicious URLs is now essential for improving network security. The technological revolution spurred ...

• We introduce a post-training program for pre-trained models.
• Character-level and subword-level URL insights are fused.
• Our method blends multi-layer Transformer encodings.
• Our method captures and integrates local and global ...

The non-parallel disease progression and curative effect are the difficulties of clinical diagnosis and treatment decisions. Experts (doctors) constantly summarize these non-parallel phenomena for more accurate diagnosis and treatment. In order ...

• A novel approach to investigate the consistency of multi-modal data-driven clinical decision making.
• Using DRSA to explore clinical non-parallel decision-making problems.
• Explore consistent decision-making methods in clinical non-...

Deep learning, especially convolutional neural networks (CNNs) and Transformer architectures, have become the focus of extensive research in medical image segmentation, achieving impressive results. However, CNNs come with inductive biases that ...

• A novel U-shaped encoder–decoder network architecture CSWin-UNet is designed for medical image segmentation.
• A Transformer-based mechanism was incorporated to implement horizontal and vertical stripes self-attention learning.
• A ...

Multi-view clustering is widely applied in engineering and scientific research. It helps reveal the underlying structures and correlations behind complex multi-view data. Graph-based multi-view clustering stands as a prominent research frontier ...

• Multi-view High-order Bipartite Graph Fusion for Enhanced Clustering.
• Adaptive Fusion with Truncation Selection and Implicit Weighting
• Computational Efficiency and Superiority in Multi-View Clustering.

• An underwater exposure adjustment method using bio-vision is proposed.
• An underwater color correction method using bio-visual perception is proposed.
• A physical model for visibility recovery using underwater features is proposed.

Intelligent monitoring of deep-sea nets is affected by light attenuation, light scattering, and limited dynamic range factors of the camera, which can cause color shift, low visibility, and over/underexposure in monitoring, resulting in reduced ...

High-precision fault diagnosis of rotating machinery plays an important role in industrial systems. Today, rotating machinery often has multiple sensors to monitor equipment condition, so it is important to fuse data from multiple rotating ...

• A multi-layer graph data construction method is proposed.
• A multi-type feature fusion mechanism based on the attention mechanism is proposed.
• A decision fusion strategy based on information entropy is proposed.
• A multi-level ...

Federated learning is a privacy-preserving distributed framework that facilitates information fusion and sharing among different clients, enabling the training of a global model without exposing raw data. However, the gradient inversion attack ...

• An effective gradient inversion attack with Diffusion Models is proposed.
• The proposed group consistency regularization can address the issue of spatial variations.
• Diffusion models as prior knowledge is employed for effective ...

Rapid and accurate diagnosis of mental disorders has long been an essential challenge in clinical medicine. Due to the advantage in addressing non-Euclidean structures, graph neural networks have been increasingly used to study brain networks. ...

• This study proposes a novel approach (FC-HGNN) for mental disorder identification.
• The FC-HGNN combines the advantages of individual and population graph models.
• LGP module was used to enhance interpretability in identifying ...

As a core branch of financial forecasting, stock forecasting plays a crucial role for financial analysts, investors, and policymakers in managing risks and optimizing investment strategies, significantly enhancing the efficiency and effectiveness ...

• Reviews the literature on data-driven neural networks in the field of stock forecasting.
• Outlines the commonly used datasets and various evaluation metrics in the field of stock forecasting.
• Explores unresolved issues and potential ...

Disease assessment involves two stages of decision-making process. The first process is to infer the development trend of the disease through the existing judgment conditions, which is to make a positive estimation (prediction) decision. The ...

• Three-way decision-making and semi-supervised learning integrated model was constructed.
• The multi-modal decision information method was discussed.
• Automatic update strategy of thresholds and loss function of three-way decision ...

Recently, many works have been proposed for traffic forecasting to improve people’s daily lives. Although these works have achieved good predictive performance, they have three fundamental limitations. (i) The regional features of traffic flow ...

• Designing a hierarchical graph learning model for traffic forecasting.
• Using Neural Ordinary Differential Equations to model continuous traffic features.
• Adopting a spatio-temporal feature fusion scheme for traffic regions and ...

The brain basis of emotion has consistently received widespread attention, attracting a large number of studies to explore this cutting-edge topic. However, the methods employed in these studies typically only model the pairwise relationship ...

• Identify hierarchical emotional areas to study brain mechanisms underlying emotion.
• Conduct an in-depth theoretical analysis based on information fusion and graph theory.
• Develop a novel framework to improve emotion decoding using ...

Animal re-identification is fundamental in the ecology and ethology community for understanding the Earth’s ecosystems. Due to the uncertainty of photographing animals in the wild, such as the position and angle of capture and the variations in ...

• For the first time, a novel hierarchical spatial-frequency transformer is proposed.
• k-NN-based self- and cross- attentions can promote both local and global information.
• The proposed fusion transformer with good performance can ...