Egocentric visual query localization (EgoVQL) focuses on localizing the target of interest in space and time from first-person videos, given a visual query. Despite recent progressive, existing methods often struggle to handle severe object appearance changes and cluttering background in the video due to lacking sufficient target cues, leading to degradation. Addressing this, we introduce PRVQL, a novel Progressive knowledge-guided Refinement framework for EgoVQL. The core is to continuously exploit target-relevant knowledge directly from videos and utilize it as guidance to refine both query and video features for improving target localization. Our PRVQL contains multiple processing stages. The target knowledge from one stage, comprising appearance and spatial knowledge extracted via two specially designed knowledge learning modules, are utilized as guidance to refine the query and videos features for the next stage, which are used to generate more accurate knowledge for further feature refinement. With such a progressive process, target knowledge in PRVQL can be gradually improved, which, in turn, leads to better refined query and video features for localization in the final stage. Compared to previous methods, our PRVQL, besides the given object cues, enjoys additional crucial target information from a video as guidance to refine features, and hence enhances EgoVQL in complicated scenes. In our experiments on challenging Ego4D, PRVQL achieves state-of-the-art result and largely surpasses other methods, showing its efficacy. Our code, model and results will be released at https://github.com/fb-reps/PRVQL.

Faces synthesized by diffusion models (DMs) with high-quality and controllable attributes pose a significant challenge for Deepfake detection. Most state-of-the-art detectors only yield a binary decision, incapable of forgery localization, attribution of forgery methods, and providing analysis on the cause of forgeries. In this work, we integrate Multimodal Large Language Models (MLLMs) within DM-based face forensics, and propose a fine-grained analysis triad framework called VLForgery, that can 1) predict falsified facial images; 2) locate the falsified face regions subjected to partial synthesis; and 3) attribute the synthesis with specific generators. To achieve the above goals, we introduce VLF (Visual Language Forensics), a novel and diverse synthesis face dataset designed to facilitate rich interactions between `Visual' and `Language' modalities in MLLMs. Additionally, we propose an extrinsic knowledge-guided description method, termed EkCot, which leverages knowledge from the image generation pipeline to enable MLLMs to quickly capture image content. Furthermore, we introduce a low-level vision comparison pipeline designed to identify differential features between real and fake that MLLMs can inherently understand. These features are then incorporated into EkCot, enhancing its ability to analyze forgeries in a structured manner, following the sequence of detection, localization, and attribution. Extensive experiments demonstrate that VLForgery outperforms other state-of-the-art forensic approaches in detection accuracy, with additional potential for falsified region localization and attribution analysis.

Establishing object-level correspondence between egocentric and exocentric views is essential for intelligent assistants to deliver precise and intuitive visual guidance. However, this task faces numerous challenges, including extreme viewpoint variations, occlusions, and the presence of small objects. Existing approaches usually borrow solutions from video object segmentation models, but still suffer from the aforementioned challenges. Recently, the Segment Anything Model 2 (SAM 2) has shown strong generalization capabilities and excellent performance in video object segmentation. Yet, when simply applied to the ego-exo correspondence (EEC) task, SAM 2 encounters severe difficulties due to ineffective ego-exo feature fusion and limited long-term memory capacity, especially for long videos. Addressing these problems, we propose a novel EEC framework based on SAM 2 with long-term memories by presenting a dual-memory architecture and an adaptive feature routing module inspired by Mixture-of-Experts (MoE). Compared to SAM 2, our approach features **(i)** a Memory-View MoE module which consists of a dual-branch routing mechanism to adaptively assign contribution weights to each expert feature along both channel and spatial dimensions, and **(ii)** a dual-memory bank system with a simple yet effective compression strategy to retain critical long-term information while eliminating redundancy. In the extensive experiments on the challenging EgoExo4D benchmark, our method, dubbed ***LM-EEC***, achieves new state-of-the-art results and significantly outperforms existing methods and the SAM 2 baseline, showcasing its strong generalization across diverse scenarios. Our code and model are available at https://github.com/juneyeeHu/LM-EEC.

Current AIGC detectors often achieve near-perfect accuracy on images produced by the same generator used for training but struggle to generalize to outputs from unseen generators. We trace this failure in part to latent prior bias: detectors learn shortcuts tied to patterns stemming from the initial noise vector rather than learning robust generative artifacts. To address this, we propose \textbf{On-Manifold Adversarial Training (OMAT)}: by optimizing the initial latent noise of diffusion models under fixed conditioning, we generate \emph{on-manifold} adversarial examples that remain on the generator’s output manifold—unlike pixel-space attacks, which introduce off-manifold perturbations that the generator itself cannot reproduce and that can obscure the true discriminative artifacts. To test against state-of-the-art generative models, we introduce GenImage++, a test-only benchmark of outputs from advanced generators (Flux.1, SD3) with extended prompts and diverse styles. We apply our adversarial-training paradigm to ResNet50 and CLIP baselines and evaluate across existing AIGC forensic benchmarks and recent challenge datasets. Extensive experiments show that adversarially trained detectors significantly improve cross-generator performance without any network redesign. Our findings on latent-prior bias offer valuable insights for future dataset construction and detector evaluation, guiding the development of more robust and generalizable AIGC forensic methodologies.

DeepFake, an AI technology that can automatically synthesize facial forgeries, has recently attracted worldwide attention. While DeepFakes can be entertaining, they can also be used to spread falsified information or be weaponized as cognition warfare. Forensic researchers have been dedicated to designing defensive algorithms to combat such disinformation. However, attacking technologies have been developed to make DeepFake products more aggressive. For example, by launching anti-forensics and adversarial attacks, DeepFakes can be disguised as authentic media to evade forensic detectors. However, such manipulations often sacrifice image quality for satisfactory undetectability. To address this issue, we propose a method to generate a novel adversarial sharpening mask for launching black-box anti-forensics attacks. Unlike many existing methods, our approach injects perturbations that allow DeepFakes to achieve high anti-forensics performance while maintaining pleasant sharpening visual effects. Experimental evaluations demonstrate that our method successfully disrupts state-of-the-art DeepFake detectors. Moreover, compared to images processed by existing DeepFake anti-forensics methods, our method’s quality of anti-forensics DeepFakes rendered is significantly improved. Our code is available at https://github.com/fb-reps/HQ-AF_GAN .

DeepFake, an AI technology for creating facial forgeries, has garnered global attention. Amid such circumstances, forensics researchers focus on developing defensive algorithms to counter these threats. In contrast, there are techniques developed for enhancing the aggressiveness of DeepFake, e.g., through anti-forensics attacks, to disrupt forensic detectors. However, such attacks often sacrifice image visual quality for improved undetectability. To address this issue, we propose a method to generate novel adversarial sharpening masks for launching black-box anti-forensics attacks. Unlike many existing arts, with such perturbations injected, DeepFakes could achieve high anti-forensics performance while exhibiting pleasant sharpening visual effects. After experimental evaluations, we prove that the proposed method could successfully disrupt the state-of-the-art DeepFake detectors. Besides, compared with the images processed by existing DeepFake anti-forensics methods, the visual qualities of antiforensics DeepFakes rendered by the proposed method are significantly refined.

With the rapid development of Internet of Things (IoT) technology, the number of sensors in IoT is proliferating and the monitoring range is expanding. Wireless sensor networks (WSNs) in IoT are large-scale WSNs with fast changes in node energy due to long-distance communication, and node distribution due to the death of nodes. A clustering and routing algorithm for the fast changes (FC-CRAs) of large-scale WSNs in IoT is proposed. In the FC-CRA, the clusters are constructed according to the cluster radius which can dynamically adapt to the change in node energy and distribution. The intracluster routing is based on the path energy function to save and balance node energy. The intercluster routing is performed in an intercluster communication node set to avoid the premature death of near-BS nodes and ensure the continuity of data transmission. Compared with the other algorithms, the FC-CRA can improve the node lifetime, network throughput, and save the network energy in large-scale or sparse WSNs. The proposed algorithm can alleviate the “energy hole” problem and improve the reliability of data transmission, which is very important for some IoT application scenes with wide-area monitoring requirement and harsh or hazardous situations.

This paper proposes a dual-channel network for DeepFake detection. The network comprises two channels: one using a stacked Maxvit block to process the downsampled original images, and the other using a stacked ResNet basic block to capture features from the discrete cosine transform of the image spectrums. The components extracted from the two channels are concatenated using a linear layer to train the entire model for exposing DeepFakes. Experimental results demonstrate that the proposed method could achieve satisfactory forensics performance. Besides, the experiments of cross-dataset evaluations prove it is also high in generalizability.

Falsified faces generated by DeepFake are severe threats to our community. Many smart systems in Industry 4.0, such as electronic payments and identity verification, rely on bioinformation authentication. These applications may compromise with forgeries generated by DeepFake. Notwithstanding many promising results on DeepFake forensics have been reported recently, we are now facing new security challenges brought by antiforensics attacks. With adversarial perturbations injected by antiforensics algorithms, falsified faces could masquerade themselves to disrupt forensics detectors as well as industrial applications. Therefore, to secure biometric data, in particular facial information, we propose a countermeasure against the attacks of DeepFake antiforensics. The proposed model features dual channels and multiple supervisors to capture biological attributes from manifold aspects. After training, the proposed method can purify antiforensics images by eliminating adversarial perturbations. With experimental evaluations, we show that purified faces are highly distinguishable from real ones. The proposed method is justified as a reliable defense tool for protecting facial bioinformation against antiforensics amid Industry 4.0.

Abstract In order to maximize the work efficiency of wireless mobile charger, a payoff‐maximization‐based adaptive hierarchical wireless charging algorithm for mobile charger is proposed. Based on the mesh structure and multi‐node charging technology, the recharging optimization for massive devices is modelled as a problem of payoff maximization. According to energy allocation, anchor point deployment and time allocation, we decompose it into three layers by the hierarchical decomposition method to obtain optimal solution quickly. The process of energy allocation and anchor point deployment in each mesh is optimized in the first two layers based on Karush–Kuhn–Trucker (KKT) condition and greedy strategy, respectively. Based on the feedback of the first two layers, the most complex problem of time allocation in the last layer is solved by our innovative gain recall mechanism. The trade‐off between the number of recharged devices and recharging time in each cycle can be achieved by only charging the devices in the meshes which are without recall gains. The simulation results prove our algorithm can adaptively adjust the ratio of moving time to recharging time in a fixed cycle, and mobile charger can always work in efficient recharging positions, whose effect is exploited utmost.

Abstract The software-defined networks-enable mobile edge computing (SDN-enable MEC) architecture, which integrates SDN and MEC technologies, realizes the flexibility and dynamic management of the underlying network resources by the MEC, reduces the distance between the access terminal and computing resources and network resources, and increases the terminal's access to resources. However, the static distribution relationship between MEC servers (MECSs) and controllers in the multi-controller architecture may result in unbalanced load distribution among the controllers, which would degrade network performance. In this paper, a multi-objective optimization MECS redistribution algorithm (MOSRA) is proposed to decrease the response time and overhead. A controller response time model and link transmit overhead model are introduced as basis of an evolutionary algorithm which is proposed to optimize MECS redistribution. The proposed algorithm aims to select an available sub-optimizes individual by using a strategy based coordination transformation from Pareto Front. That is, when the master controller of the MECS is redistributed, both of the network overhead of the MECS to the controller and the response time of the controller to the MECS processing request are optimized. Finally, the simulation results demonstrate that the MOSRA can solve the redistribution problem in different network load levels and different network sizes within the effective time, and has a lower control plane response time, while making the edge network plane transmission overhead lower, compared with other algorithms .

As the support networks of the electric power grid, power communication networks (PCNs) become more complex and vulnerable due to the increasing scale of the electric power grid. Identifying and protecting the critical nodes in PCNs in advance is an effective way to reduce network vulnerability. Owing to the large differences in vulnerability indicators of different layers in the PCN, it is difficult to find the critical nodes, which have great impacts on all vulnerability indicators. Therefore, the goal of this study is to identify the critical nodes that have greater impacts on different layers, rather than nodes that have the greatest impact on a single layer. Therefore, the authors present a model to analyse the node in the PCNs from the physical topology, traffic distribution, and service importance distribution to calculate the node importance in the physical topology layer, the transport layer, and the service layer, respectively. Combined with a multi‐layer critical nodes identification algorithm (MCNIA) proposed, the node critical degree is obtained so that it can identify the critical nodes in the PCNs. The vulnerability analyses of PCNs under critical nodes attacking prove that MCNIA can identify critical nodes in the PCNs precisely.

The query image is usually a simple and single object in image retrieval, and the reference images in the database usually have many distractions. The precision of image retrieval can be greatly improved If the target regions in the database image are extracted during retrieval. So this paper proposes a Bow image retrieval method based on SSD target detection. First, the training gallery is manually annotated to record the location and size information. Second, the SSD target detection model is trained with the labeled training gallery to obtain the target object SSD model. Third, the SSD model is used to locate the similar target regions of the reference image and the query graph. Finally, the target region information is mapped into the convolutional features, and these feature vectors are used for image similarity matching. The performance of the proposed method is evaluated on Paris6k, Oxford5k, Paris106k and Oxford105k databases. The experimental results show that the accuracy of image retrieval will be greatly improved by adding optimization methods in the proposed image retrieval framework. The image retrieval accuracy of this method is higher than that of similar methods in recent years.

This paper initially presents the channel modeling for full duplex (FDX) modulating retro-reflector (MRR) freespace optical (FSO) systems under weak-to-strong turbulence conditions. The uplink channel in the MRR FSO communication is a conventional FSO channel that is typically modeled by a Gamma-Gamma (ΓΓ) distribution. The downlink channel is jointly affected by the turbulence-induced fading in two opposite passes as well as the reflection effect of the MRR, and its normalized channel fading can be described as the product of two correlated ΓΓ random variables (RVs). The channel model of the downlink is approximated by a single α-μ distribution via moment matching method. Then, we propose an effective signal modulation and detection scheme to achieve the FDX communication. The bit error rates of the MRR FSO systems are calculated by the analytical formulas that are derived based on the channel models of the two links, and also evaluated by the Monte-Carlo simulations to validate the derived formulas. The effect of system configurations and propagation conditions on the system performance is investigated, and the optimal designs to make a trade-off between the uplink and downlink performance are discussed.

The new generation intelligent park is demanded for the flexible utilization mode of distributed renewable energy, energy storage equipment and the various characteristic load access. However, the energy router is correspondingly responsible for energy routing and dynamic control. Its uniform interface supports the plug and play of distributed power, energy storage and load of the intelligent park. In our paper, the new energy network architecture are firstly presented. Then, the function structure and the bidirectional flow control strategy of these energy routers are designed to satisfy the energy routing demands. The bidirectional flow control among users and a variety of distributed power, energy storage equipment is achieved by implementing our energy exchange scheme of the energy routing interface design. Meanwhile, we put the energy router interface design into the practical operation in local energy internet, construct the application system based on the energy internet technology with Cell in intelligent park, and ultimately make the collaborative operation and green resource sharing in intelligent system.

Multitrack detection with array-head reading is a promising technique proposed for next generation magnetic storage systems. The multihead multitrack (MHMT) system is characterized by intersymbol interference in the downtrack direction and intertrack interference (ITI) in the crosstrack direction. Constructing the trellis of a MHMT maximum likelihood (ML) detector requires knowledge of the ITI, which is generally unknown at the receiver. Furthermore, in a time-varying ITI environment, updating ML trellis labels using adaptively-generated ITI estimates could incur significant delay. In this paper, we propose one approach to solve these issues. The proposed detector uses a different trellis structure whose output labels are independent of the ITI level, with ITI-dependence appearing only in a scale factor used to suitably weight the computed path metrics in order to retain ML optimality. The detector formulation facilitates the design of a gain loop structure that can track the time-varying ITI and provide ITI estimates to adaptively adjust the weights in the path metric evaluation. Simulation results show that the proposed detector architecture with ITI estimation offers a substantial performance advantage over ML detection using a static ITI estimate.

To achieve large storage capacity on magnetic hard disk drives, very high track density is required, causing severe intertrack interference (ITI). Multihead multitrack (MHMT) detection has been proposed to better combat the effects of ITI. Such detection, however, has prohibitive implementation complexity. Reduced-state sequence estimation (RSSE) is a promising technique for significantly reducing the complexity, while retaining good performance. In this paper, several different MHMT models are considered, including symmetric and asymmetric 2H2T systems, and a symmetric 3H3T system. By carefully evaluating the effective distance between two input symbols, we propose optimized set partition trees for each channel model. Different trellis configurations for RSSE are constructed based on the desired performance/complexity tradeoff. Simulation results show that the reduced MHMT detector can achieve near maximum-likelihood (ML) performance with a small fraction of the original number of trellis states. We also use error event analysis to explain the behavior of RSSE. The proposed algorithm could be potentially applied to next generation magnetic recording systems, especially when the ML detector is infeasible due to the high computational complexity.

Multitrack detection for shingled magnetic recording (SMR) using a two-head array system is considered. The channel suffers from intersymbol interference (ISI) in the down-track direction and intertrack interference (ITI) in the crosstrack direction. We propose a practical multihead/multitrack detector that provides a low-complexity approach to adaptive estimation of time-varying ITI. The performance of the proposed detection algorithm is analyzed in terms of its minimum distance parameter, and simulation results show that the proposed detector offers a performance advantage in settings where complexity constraints limit the maximum-likelihood two-track detector to use a static ITI estimate.