Search

The cell-free network is regarded as an enabling technology for alleviating inter-cell interference due to its user-centric network paradigm. Reconfigurable intelligent surfaces (RISs) are used to assist cell-free networks in improving network capacity at a low cost. However, only half-space coverage can be achieved by RISs. With the unique capability of extending half-space coverage to full-space coverage, simultaneously transmitting and reflecting RIS (STAR-RIS) is emerged as an attractive technology. Motivated by this distinctive characteristic, multiple STAR-RISs deployed in millimeter wave (mmWave) cell-free systems to assist users in downlink transmission are investigated in this paper. Our objective is to maximize the weighted sum-rate (WSR) via joint optimization of association indicators, active beamforming vectors at mmWave base stations (MBSs) as well as transmission and reflection beamforming vectors of STAR-RISs, while the limits of STAR-RIS-user association, each MBS maximum transmit power as well as passive coefficients of STAR-RISs ought to be satisfied. To handle the non-convex problem, we develop the Lagrangian dual transform (LDT)-quadratic transform (QT)-based algorithm. To be specific, the LDT and QT are utilized for reformulating the non-convex problem as a more solvable equivalent one, afterwards, the transformed problem is decomposed into five subproblems via the alternating optimization (AO) method. Particularly, the solutions of association indicators, active beamforming vectors, and passive beamforming vectors are obtained through the linear conic relaxation (LCR)-based association scheme, the rank-one relaxation approach as well as the penalty-based norm approximation algorithm, respectively. Simulation results indicate that the proposed method can obtain better WSR performance over benchmarks. Some useful guidances are also provided for mmWave cell-free systems.

Accurate and efficient keypoint detection and description is a fundamental step in various computer vision tasks. In this paper, we extract robust descriptors and detect accurate keypoints by learning local Features with Domain adaptation (DomainFeat). Specifically, our Domainfeat includes image-level domain invariance supervision, pixel-level domain consistency supervision, Pixel-Adaptive keypoint Detection(PA-Det), and cross-domain dataset with domain stable point supervision. First, we introduce the image-level domain invariance supervision to make the high-level feature distributions from different domains close by fusing domain-invariant representations in the decoder. Furthermore, to compensate for the inconsistency between descriptors corresponding to the keypoints at the pixel level, we propose the pixel-level domain consistency supervision. Then we present the Pixel-Adaptive keypoint Detection to efficiently detect accurate keypoints, which can improve accuracy by enhancing the local consistency of heatmaps. Finally, we propose an efficient approach to construct data and supervision labels in diverse domains, which can tackle complex application scenarios. With these novel modules and supervision methods, our DomainFeat can make feature detectors more accurate and descriptors more robust. Extensive experiments confirm that Domainfeat achieves state-of-the-art performance on benchmarks such as Aachen-Day-Night localization, HPatches image matching, and the challenging DNIM dataset.

In this article, an efficient federated learning (FL) Framework in the Internet of Vehicles (IoV) is studied. In the considered model, vehicle users implement an FL algorithm by training their local FL models and sending their models to a base station (BS) that generates a global FL model through the model aggregation. Since each user owns data samples with diverse sizes and different quality, it is necessary for the BS to select the proper participating users to acquire a better global model. Meanwhile, considering the high computational overhead of existing selection methods based on the gradient, the lightweight user selection scheme based on the loss decay is proposed. Due to the limited wireless bandwidth, the BS needs to select an suitable subset of users to implement the FL algorithm. Moreover, the vehicle users' computing resource that can be used for FL training is usually limited in the IoV when other multiple tasks are required to be executed. The local model training and model parameter transmission of FL will have significant effects on the latency of FL. To address this issue, the joint communication and computing optimization problem is formulated whose objective is to minimize the FL delay in the resource-constrained system. To solve the complex nonconvex problem, an algorithm based on the concave-convex procedure (CCCP) is proposed, which can achieve superior performance in the small-scale and delay-insensitive FL system. Due to the fact that the convergence rate of CCCP method is too slow in a large-scale FL system, this method is not suitable for delay-sensitive applications. To solve this issue, a block coordinate descent algorithm based on the one-step projected gradient method is proposed to decrease the complexity of the solution at the cost of light performance degrading. Simulations are conducted and numerical results show the good performance of the proposed methods.

Local features detection and description are widely used in many vision applications with high industrial and commercial demands. With large-scale applications, these tasks raise high expectations for both the accuracy and speed of local features. Most existing studies on local features learning focus on the local descriptions of individual keypoints, which neglect their relationships established from global spatial awareness. In this paper, we present AWDesc with a consistent attention mechanism (CoAM) that opens up the possibility for local descriptors to embrace image-level spatial awareness in both the training and matching stages. For local features detection, we adopt local features detection with feature pyramid to obtain more stable and accurate keypoints localization. For local features description, we provide two versions of AWDesc to cope with different accuracy and speed requirements. On the one hand, we introduce Context Augmentation to address the inherent locality of convolutional neural networks by injecting non-local context information, so that local descriptors can “look wider to describe better”. Specifically, well-designed Adaptive Global Context Augmented Module (AGCA) and Diverse Surrounding Context Augmented Module (DSCA) are proposed to construct robust local descriptors with context information from global to surrounding. On the other hand, we design an extremely lightweight backbone network coupled with the proposed special knowledge distillation strategy to achieve the best trade-off in accuracy and speed. What is more, we perform thorough experiments on image matching, homography estimation, visual localization, and 3D reconstruction tasks, and the results demonstrate that our method surpasses the current state-of-the-art local descriptors. Code is available at: https://github.com/vignywang/AWDesc.

Automatically extracting roads from very high-resolution (VHR) remote sensing images is of great importance in a wide range of remote sensing applications. However, complex shapes of roads (i.e., long, geometrically deformed, and thin) always affected the extraction accuracy, which is one of the challenges of road extraction. Based on the insight into road shape characteristics, we propose a novel road shape-aware network (RSANet) to achieve efficient and accurate road extraction. First, we introduce the efficient strip transformer module (ESTM) to efficiently capture the global context to model the long-distance dependence required by long roads. Second, we design a geometric deformation estimation module (GDEM) to adaptively extract the context from the shape deformation caused by shooting roads from different perspectives. Third, we provide a simple but effective road edge focal loss (REF loss) to make the network focus on optimizing the pixels around the road to alleviate the unbalanced distribution of foreground and background pixels caused by the roads being too thin. Finally, we conduct extensive evaluations on public datasets to verify the effectiveness of RSANet and each of the proposed components. Experiments validate that our RSANet outperforms state-of-the-art methods for road extraction in remote sensing images.