This paper proposes a secure and energy-efficient cluster-based V2V communication framework that integrates a Proof of Authority (PoA) consensus mechanism, Verifiable Delay Function (VDF)-based Cluster Head (CH) selection, and Shamir’s Secret Sharing (SSS). The VDF guarantees unbiased, verifiable, and tamper-resistant selection of CHs. Each CH is validated through a decentralized mechanism and entrusted with maintaining the blockchain ledger. SSS divides each CH’s private key among neighboring nodes to eliminate single points of failure (SPF), ensuring resilience. Elliptic curve cryptography ensures lightweight encryption, while a cluster-based architecture reduces communication overhead and enhances scalability. Simulation results demonstrate a 58% reduction in latency, a 54% improvement in energy efficiency, and a 22% increase in packet delivery ratio compared to existing schemes such as ECT-POW, CBT-PoS, and VPKI-DAG. The proposed system also achieves higher security robustness through verifiable leader election, key fault tolerance, and authenticated communication channels.

Abstract Knowledge tracing (KT) refers to the problem of predicting a learner’s future performance based on their past performance in education. Recently, attention-based sequence modeling methods achieve impressive predictive performance. However, existing solutions merely consider one single sequence modeling method, which might fail to capture the comprehensive state of knowledge across long sequences. In this paper, we propose D ual S equence M odeling for K nowledge T racing (DSMKT). DSMKT aims to enhance the modeling of a learner’s long-term profile by collaborating two sequence modeling methods, i.e., the masked self-attention mechanism and the gated recurrent unit. To further exploit the synergy between two sequence models, we adopt the idea of online knowledge distillation and adaptively combine two branches to form a stronger teacher model, which in turn provides predictions as extra supervision for better modeling ability. Extensive experiments on four real-world benchmark datasets show that DSMKT performs excellently in predicting future learner responses.

The rapid development of Internet of Things (IoT) applications has imposed stringent demands on low-latency and energy-efficient task offloading. Although mobile edge computing (MEC) provides nearby computing capabilities, conventional data-driven offloading approaches suffer from redundant transmission and limited scalability under constrained wireless and computational resources. To address this, we propose a general semantic-aware offloading framework that integrates semantic communication, Non-Orthogonal Multiple Access (NOMA), MEC, and user mobility prediction in edge-enabled IoT networks. The offloading user first performs local semantic extraction based on a tunable semantic factor that influences both the local computation cost and the transmitted data volume, and theoretical modeling focuses on adjustable granularity to derive an approximate optimal theoretical bound. The resulting semantic information is then transmitted to the MEC server for task execution. We formulate a joint optimization problem to minimize total user-side energy consumption by optimizing offloading decisions, semantic factors, transmit power, and time-slot scheduling. To solve this, we develop a Mixed-Integer Linear Programming (MILP)-based solution via linearization and piecewise approximation of the original non-convex problem, and we further propose a low-complexity heuristic algorithm for practical feasibility.

The integration of Digital Twin (DT) technology with Mobile Edge Computing (MEC) offers a promising solution for real-time system monitoring and optimization in smart industrial environments. To fully exploit the advantages of their integration, the allocation of MEC resources should carefully consider the application requirements of DT scenarios. This paper focuses on MEC-assisted DT scenarios and proposes a framework to model the dynamic interaction between device states, synchronization demands, and resource allocation, considering the varying synchronization frequency required by different device operational states. Under constraints of communication, computation, and storage resources, we perform joint optimization of the synchronization frequency and task offloading strategies. From both theoretical analysis and practical application perspectives, an optimal strategy based on solving an Integer Linear Programming (ILP) problem and a low-complexity heuristic algorithm are proposed. Extensive simulations in a smart factory scenario demonstrate significant advantages in improving the value of the utility function. This paper provides new insights into the deep integration of MEC and DT technologies and offers valuable references for applications in resource-constrained industrial environments.

Convolutional Neural Networks and Attention-based Transformer have emerged as the preferred models for medical image processing. Recently, specific network architectures relying solely on multilayer perceptrons (MLPs) have gained popularity and demonstrated excellent results in various computer vision tasks. In particular, CycleMLP has demonstrated good performance in dense prediction tasks owing to its adaptability to image size and linear computational complexity. However, the basic operator of CycleMLP has a fixed sampling location for any feature map and samples very few target organs in medical images characterized by an extreme imbalance between foreground and background. Therefore, effectively extracting the features of target organs becomes challenging. In this paper, we propose a new MLP-like module, SlideMLP, by considering the sparsity of target organs in medical images. This module extracts a set of offsets from the input feature maps and utilizes these offsets to re-select the sampling points. This approach effectively enhances the sampling rate of target organs while retaining the advantages of CycleMLP. Additionally, we constructed a U-shaped network with a pure MLP using this module and assessed its robustness using two datasets with different modalities. Comparative results with state-of-the-art methods demonstrate that the method proposed in this paper can achieve a substantial Dice Similarity Coe cient (DSC) while utilizing fewer parameters.

Hybrid non-orthogonal multiple access (NOMA) is a promising scheme to significantly improve communication performance within a 5G mobile network. In this article, a NOMA network is established in the context of underlay cognitive radio (CR) firstly. In addition, the combination of artificial noise (AN) and decode-and-forward (DF) technology is applied at the unmanned aerial vehicle (UAV) relay. We then derive the secrecy outage probability (SOP), which is affected by power allocation factors, to measure the system secrecy performance. Finally, a machine learning (ML) power allocation optimization algorithm is proposed considering both network security and user fairness. To be specific, we employ a revised decision tree (DT) embedded multilayer perceptron (MLP) in the leaf nodes, aiming to utilize the strengths of both ML algorithms effectively and find the optimal groups of systems with various parameters rapidly. The data results demonstrate that the introduced algorithm provides satisfactory predictions and obtains more accurate coefficients compared to other ML algorithms.

As an indispensable part of the future 6G communication system, SIoT (Satellite Internet of Things) plays a vital role in global communication. However, the communication requirements of users, the manufacturing and launching costs of satellites, and the performance coverage of constellations pose significant challenges to the construction of SIoT. To facilitate the evolution of the 5G network into the 6G space-ground integrated network, this paper proposes a multi-area on-demand classification (MOC) constellation design based on CubeSats. Firstly, the area of interest is classified into two categories: the area with cellular base stations and the area without cellular base stations. Secondly, three coverage, communication quality, and cost models are established. The coverage model establishes the initial configuration of the constellation and determines the value for evaluating coverage. The communication model determines the evaluation values for communication quality. The cost model determines the evaluation values for cost. Finally, the multi-objective Manta ray foraging optimization algorithm (MOMRFO) optimizes the three evaluation values to obtain the Pareto front for the best configuration constellation. Based on actual data from the cellular base station, the simulation results have demonstrated the MOC constellation scheme’s feasibility and effectiveness for future SIoT deployment.

Despite rapid progress of end-to-end optimization for single-image dehazing, a long-standing open problem is the non-homogenous haze, at the core of the differences between synthetic hazy images and real hazy images. The atmospheric scattering model (ASM) has been widely adopted to model the degradation process of haze images but based on the assumption of homogeneous haze. In realistic scenarios, non-homogeneous haze often makes it more difficult to estimate the transmission map in ASM, resulting in undesired artifacts in the restored images. To address the issue of non-homogeneous haze, we propose to model the uncertainty in the estimation of the transmission map and develop a spatially adaptive learning module for ASM correction. Specifically, we present an approach to enhancing the well-known Dark Channel prior (DCP) by relaxing the constraint with the transmission map in the DCP-net. Assuming the availability of paired training data, we have developed a strategy to address vulnerability in the DCP, leading to a more accurate estimation of the transmission map. Then, we explore the uncertainty between the estimated transmission map and target transmission map (Ground Truth) to reformulate the ASM for the presence of non-homogeneous haze. A robust and accurate estimated transmission map can boost the final dehazing performance of our DCP-net. Experiments on three popular synthetic and real non-homogeneous datasets show that our proposed approach has achieved better results on both synthetic scenes and real non-homogeneous scenes. The code is available athttps://see.xidian.edu.cn/faculty/wsdong/Projects/Projects/project_dehazing_TCSVT2024.htm

Edge caching is a widely adopted technique for improving network performance and user experience. To optimize its benefits, researchers are exploring the use of recommendation systems, which can leverage advanced algorithms to determine which content should be cached at the edge, leading to lower latency, reduced bandwidth usage, and ultimately better network and service management. The majority of existing works, on the other hand, are based on independently and identically distributed request patterns. In this work, the problem of joint caching and recommendation policy for long-term user request behavior was studied, which is more realistic. Specifically, the continuous request behavior of users was regarded as being relevant and was modeled as an Absorbing Markov Chain. The goal was to minimize the expected content delivery cost over multiple viewing sessions while meeting the requirements for expected value on quality of recommendation. The optimization problem was formulated and transformed into a mixed-integer programming (MIP) problem, which provides an optimal solution (JCRP). Meanwhile, a heuristic algorithm with low computational complexity that is more friendly to network resources has been proposed (LvJCR). Simulation results show that the proposed algorithms have advantages in cost minimization over long sessions.

Blind face restoration aims to recover high-quality face images from low-quality ones with complex and unknown degradation. Existing approaches have achieved promising performance by leveraging pre-trained dictionaries or generative priors. However, these methods may fail to exploit the full potential of degraded inputs and facial identity features due to complex degradation. To address this issue, we propose a novel method that explores the correlation of degraded spatial identity features by learning a general representation using memory network. Specifically, our approach enhances degraded features with more identity by leveraging similar facial features retrieved from memory network. We also propose a fusion approach that fuses memorized spatial features with GAN prior features via affine transformation and blending fusion to improve fidelity and realism. Additionally, the memory network is updated online in an unsupervised manner along with other modules, which obviates the requirement for pre-training. Experimental results on synthetic and popular real-world datasets demonstrate the effectiveness of our proposed method, which achieves at least comparable and often better performance than other state-of-the-art approaches.

It is challenging to synchronize switched time-delay systems when some modes are uncontrolled and the dwell time (DT) of controlled mode is very small. Therefore, in this article, global exponential synchronization almost surely (GES a.s.) in a cluster of switched neural networks (NNs) with hybrid delays (time-varying delay and infinite-time distributed delay) is investigated, where transition probability (TP)-based random mode-dependent average DT (MDADT) switching is considered. A novel mode-dependent pinning event-triggered controller with nonidentical deception attacks is proposed to save the communication resource and derive less conservative results. The two necessary and restrictive conditions in existing papers that the value of the Lyapunov-Krasovskii functional (LKF) before switching instants should be smaller than that after corresponding instant and the DT of each switching mode is restricted by the sampling intervals of the event trigger are moved. Sufficient conditions in terms of linear matrix inequalities (LMIs) are given to guarantee the GES a.s., even though both synchronizing and nonsynchronizing modes coexist and maybe the minimum DT of synchronizing modes is very small. Numerical examples, including image encryption, are provided to demonstrate the merits of the new technique.

Unlike the success of neural architecture search (NAS) in high-level vision tasks, it remains challenging to find computationally efficient and memory-efficient solutions to low-level vision problems such as image restoration through NAS. One of the fundamental barriers to differential NAS-based image restoration is the optimization gap between the super-network and the sub-architectures, causing instability during the searching process. In this paper, we present a novel approach to fill this gap in image denoising application by connecting model-guided design (MoD) with NAS (MoD-NAS). Specifically, we propose to construct a new search space under a model-guided framework and develop more stable and efficient differential search strategies. MoD-NAS employs a highly reusable width search strategy and a densely connected search block to automatically select the operations of each layer as well as network width and depth via gradient descent. During the search process, the proposed MoD-NAS remains stable because of the smoother search space designed under the model-guided framework. Experimental results on several popular datasets show that our MoD-NAS method has achieved at least comparable even better PSNR performance than current state-of-the-art methods with fewer parameters, fewer flops, and less testing time. "The code associate with this paper is available at: https://see.xidian.edu.cn/faculty/wsdong/Projects/Mod-NAS.htm".

Acquiring degraded images with paired high-resolution (HR) images is often challenging, impeding the advance of image super-resolution in real-world applications. By generating realistic low-resolution (LR) images with degradation similar to that in real-world scenarios, simulated paired LR-HR data can be constructed for supervised training. However, most of the existing work ignores the degradation uncertainty of the generated realistic LR images, since only one LR image has been generated given an HR image. To address this weakness, we propose learning the degradation uncertainty of generated LR images and sampling multiple LR images from the learned LR image (mean) and degradation uncertainty (variance) and construct LR-HR pairs to train the super-resolution (SR) networks. Specifically, uncertainty can be learned by minimizing the proposed loss based on Kullback-Leibler (KL) divergence. Furthermore, the uncertainty in the feature domain is exploited by a novel perceptual loss; and we propose to calculate the adversarial loss from the gradient information in the SR stage for stable training performance and better visual quality. Experimental results on popular real-world datasets show that our proposed method has performed better than other unsupervised approaches.

Separating the background and foreground components from video frames is important to many tasks in computer vision and multimedia. As of today, robust principal component analysis (RPCA) has shown highly promising performance with the assumption that the background is low-rank and the foreground is sparse. However, existing RPCA-based methods have overlooked the uncertainty that some parts of the background (e.g., moving leaves in a dynamic background) or even the whole background (e.g., camera jittering) can be moving, which violates the low-rank assumption. To address this issue, we propose a novel enhanced RPCA framework (called ERPCA) by robustly modeling the dynamic background. Different from traditional RPCA framework, the background is decomposed into a low-rank component and a sparse component in the proposed ERPCA framework. Specifically, the sparse parts including foreground and dynamic parts of the background are modeled by Gaussian scale mixture (GSM) model. Moreover, those sparse components are further constrained by temporal consistency using nonzeromeans Gaussian models; the correspondences between sparse pixels in adjacent frames are explored by optical flow. Experimental results on 40 real videos demonstrate the superiority of our proposed method, with better average results than current state-of-the-art foreground estimation methods.

In low-level vision such as single image super-resolution (SISR), traditional MSE or L1 loss function treats every pixel equally with the assumption that the importance of all pixels is the same. However, it has been long recognized that texture and edge areas carry more important visual information than smooth areas in photographic images. How to achieve such spatial adaptation in a principled manner has been an open problem in both traditional model-based and modern learning-based approaches toward SISR. In this paper, we propose a new adaptive weighted loss for SISR to train deep networks focusing on challenging situations such as textured and edge pixels with high uncertainty. Specifically, we introduce variance estimation characterizing the uncertainty on a pixel-by-pixel basis into SISR solutions so the targeted pixels in a high-resolution image (mean) and their corresponding uncertainty (variance) can be learned simultaneously. Moreover, uncertainty estimation allows us to leverage conventional wisdom such as sparsity prior for regularizing SISR solutions. Ultimately, pixels with large certainty (e.g., texture and edge pixels) will be prioritized for SISR according to their importance to visual quality. For the first time, we demonstrate that such uncertainty-driven loss can achieve better results than MSE or L1 loss for a wide range of network architectures. Experimental results on three popular SISR networks show that our proposed uncertainty-driven loss has achieved better PSNR performance than traditional loss functions without any increased computation during testing. The code is available at https://see.xidian.edu.cn/faculty/wsdong/Projects/UDL-SR.htm

Subtracting the backgrounds from the video frames is an important step for many video analysis applications. Assuming that the backgrounds are low-rank and the foregrounds are sparse, the robust principle component analysis (RPCA)-based methods have shown promising results. However, the RPCA-based methods suffered from the scale issue, i.e., the ℓ1-sparsity regularizer fails to model the varying sparsity of the moving objects. While several efforts have been made to address this issue with advanced sparse models, previous methods cannot fully exploit the spatial-temporal correlations among the foregrounds. In this paper, we proposed a novel spatial-temporal Gaussian scale mixture (STGSM) model for foreground estimation. In the proposed STGSM model, a temporal consistent constraint is imposed over the estimated foregrounds through nonzero-means Gaussian models. Specifically, the estimates of the foregrounds obtained in the previous frame are used as the prior for these of the current frame, and nonzero means Gaussian scale mixture models (GSM) are developed. To better characterize the temporal correlations, the optical flow has been used to model the correspondences between foreground pixels in adjacent frames. The spatial correlations have also been exploited by considering that local correlated pixels should be characterized by the same STGSM model, leading to further performance improvements. Experimental results on real video datasets show that the proposed method performs comparably or even better than current state-of-the-art background subtraction methods.