The advent of the 5G RedCap, the upcoming 6G and the proliferation of the Internet of Things (IoT) have catalyzed the rapid advancement of unmanned aerial vehicle (UAV) technology while also promoting UAVs' widespread application. In IoT-enabled environments where the global positioning system (GPS) signals are compromised, visual simultaneous localization and mapping (V-SLAM) technology has emerged as an effective positioning solution, valued for its reliability. However, the presence of dynamic elements in complex environments, such as pedestrians and vehicles, poses challenges to the positioning accuracy of UAVs employing V-SLAM for navigation. This paper proposes a dynamic feature filtering-based SLAM (DFF-SLAM) approach to eliminate the impact of dynamic factors in dynamic environments, thereby enhancing the positioning accuracy of UAVs in IoT-enabled complex environments. Firstly, a semantic detection thread is designed to identify semantic information in the scene and acquire prior dynamic targets, facilitating the filtering of prior dynamic feature points. Secondly, optical flow tracking conducted at each level of the image pyramid facilitates feature point matching across consecutive images. Finally, the epipolar geometry constraint is utilized to determine the motion status of remaining feature points, further filtering out dynamic feature points. Simulation results demonstrate that compared to traditional visual SLAM systems, the UAV equipped with the DFF-SLAM system achieves more accurate positioning and meets real-time positioning requirements when navigating through IoT enabled complex environments

Total anomalous pulmonary venous connections (TAPVC) is a serious congenital heart disease, and surgery is the main treatment for TAPVC patients. Surgical repair of TAPVC is challenging, and recently, 3D visualization including 3D printing and virtual reality has been adopted in clinical practice to mitigate this problem. However, the whole heart and great vessels for 3D printing is manually segmented by experts, which is time-consuming, subject, and costly. Though whole heart and great vessel segmentation have been a hot topic in the community for decades, the existing works focus on general diseases, and There is no specific attention on TAPVC. In this paper, we collect the first whole heart and great vessel segmentation dataset for surgical planning of TAPVC. The dataset contains 92 samples, which is annotated by experienced radiologist with about 1 hour. Unlike existing works which doesn’t consider pulmonary veins and left atrium separately, 10 anatomies including left ventricle, right ventricle, left atrium, right atrium, aorta, pulmonary artery, myocardium, superior vena cava, inferior vena cava, and pulmonary veins are annotated. Then, we proposed attention ensemble based network, AE-Network for whole heart and great vessel segmentation dataset for surgical planning of TAPVC. AE-Network adopts a three-branch structure, where each branch network is based on the U-shaped network with attention mechanisms. For each of the three branch, spatial attention, channel attention, and pixel attention are used to fully exploit the context. Dynamic segmentation result fusion (DSRF) is then used to combine all the three branches to achieve accurate and robust segmentation. Experimental results show that AE-Network achieved an average Dice coefficient of 80.60%, significantly outperforming existing works. Our dataset and code will be open-sourced once the paper is accepted.

As video transmission systems expand into various complex scenarios, real-time video coding methods are essential for maintaining low latency and high perceptual quality across varying network conditions. In this work, we propose service-aware Web real-time communication (WebRTC), a semantic-assisted WebRTC system built on scalable video coding (SVC). Specifically, this system is structured with three layers: 1)$\mathcal {L}_{1}$extracts and down-samples semantic information at the encoder, employing a novel super-resolution (SR) method named BUS-DDIM at the decoder to enhance the transmission efficiency and machine vision recognition rate; 2)$\mathcal {L}_{2}$adaptively compresses high-quality video by discarding frames with little motion at the encoder to address latency issues under poor network conditions, and utilize the adjacent frame-guided denoised interpolation model called the adjacent frame-guided denoised diffusion implicit model for restoring the video; and 3)$\mathcal {L}_{3}$transmits high-quality video tailored for users with high-definition video requirements and favorable network conditions. These layers dynamically enhance the visual experience and ensure low latency across various network environments. Experiments are conducted on diverse videos to validate the effectiveness of the proposed framework. The performance evaluation under real-time scenarios indicates significant enhancements in video quality and transmission efficiency, showcasing compatibility and versatility across various applications.

Breast ultrasound (BUS) is an effective imaging modality for breast cancer diagnosis. The structural characteristics of breast lesions play an important role in computer-aided diagnosis. In this paper, a novel structure-aware triplet path network (SATPN) was designed to integrate classification and image reconstruction tasks to achieve accurate diagnosis on BUS images. Specifically, we enhanced clinically-approved structure characteristics of breast lesion by converting original BUS images to BI-RADS-oriented feature maps (BFMs) with a distance-transformation coupled Gaussian filter. Then, the converted BFMs were used as the inputs of the SATPN, which performed a supervised lesion classification task and two separate unsupervised stacked convolutional auto-encoder tasks for benign and malignant image reconstruction. We trained the SATPN with an alternative learning strategy by balancing image reconstruction error and classification label prediction error. The lesion label was determined by weighted voting of reconstruction error and label prediction error. We compared the performance of the SATPN with five deep learning methods using the original images and BFMs as inputs. Experimental results on two BUS datasets showed that SATPN performed the best among the six networks, with classification accuracy around 96%. These findings indicate that SATPN is promising for effective ultrasound computer-aided diagnosis of breast lesions.

In the era of Internet of Things (IoT), various computation-intensive applications emerge and bring great challenges to IoT devices with limited computation capability. Mobile edge computing (MEC) provides rich computing resources for IoT devices and improves applications’ execution efficiency. In this paper, we model applications as directed acyclic graphs (DAG) and target to minimize applications’ execution latency in heterogeneous MEC networks. To solve this problem, a Deep Q-Network (DQN)-based task scheduling (DQNTS) algorithm is proposed by utilizing deep reinforcement learning (DRL). By modeling the task scheduling process as a Markov decision process (MDP) and designing its critical elements, satisfying scheduling decisions are obtained. Simulation results show that the proposed algorithm achieves lower execution latency than the compared algorithms and it is adaptable to different MEC network environments.

In this paper, we investigate the joint user association and resource allocation problem in the downlink of cloud radio access networks (CRAN). The central cloud performs the scheduling of spectrum resource across different base stations (BSs) to user equipments (UEs) for the maximal network utility under practical network constraints. Nevertheless, the problem is a combinatorial optimization problem, which is intractable by the traditional exhaustive search when the network size is large. We propose a new scheduling selection method by introducing Markov approximation, which synthesizes algorithm to achieve approximate optimization by forming a reversible continuous-time Markov chain. This method adapts to not only the static network situation but also the dynamic network situation caused by users joining or departure. Besides, this method solves the two network subproblems of user association and resource allocation concurrently, which reduces the computing complexity. Simulation results demonstrate the convergence of Markov approximation and the validity of the proposed algorithm.

To satisfy the high requirements on operation efficiency in the 5G network, self-organizing network (SON) is envisioned to reduce the network operating complexity and costs by providing SON functions, which can optimize the network autonomously. However, different SON functions have different time scales and inconsistent objectives, which leads to conflicting operations and network performance degradation, raising the needs for SON coordination solutions. In this paper, we devise a multiple time-scale coordination management scheme (MTCS) for densely deployed SONs, considering the specific time scales of different SON functions. Specifically, we propose a novel analytical model named M time-scale Markov decision process, where SON decisions made in each time-scale consider the impacts of SON decisions in other M - 1 time scales on the network. Furthermore, in order to manage the network more autonomously and efficiently, a Q-learning algorithm for SON functions in the proposed MTCS scheme is proposed to achieve a stable control policy by learning from history experience. To improve energy efficiency, we then evaluate the proposed MTCS scheme with two functions of mobility load balancing and energy saving management with designed network utility. The simulation results show that the proposed SON coordination scheme significantly improves the network utility with different quality of experience requirements while guaranteeing stable operations in wireless networks.

In 5G networks, self-organizing network (SON) is envisioned to reduce the network operating complexity and costs by providing SON functions, especially in ultra- dense small cell networks. However, diverse SON functions have different time scales and inconsistent targets, which leads to the operation conflicts and network performance degradation. In this paper, we devise a multiple time-scale coordination management scheme (MTCS) to guarantee efficient and stable network operations for densely deployed SONs, where different SON functions have their own specific time scales. Specifically, we propose a novel analysis model , named M time-scale Markov decision process (MMDP), where SON decisions made in each time-scale considers the impacts of SON decisions in other M-1 time scales on the network. Then the proposed scheme with two functions of mobility load balancing (MLB) and energy saving management (ESM) is evaluated in terms of the designed network utility. Simulation results demonstrate that the proposed SON function coordination scheme significantly improves the network utility, while guaranteeing the stability of cooperative operations in wireless networks.

Small cell networks (SCNs) are envisioned as a promising solution to increase the network capacity and coverage. The densely deployments of SCNs in 5G networks pose new challenges for energy-efficient network management. Energy harvesting technique is put forward as a relatively new energy saving concept. However, due to the opportunistic nature of energy harvesting, the uncertainty and complexity will be introduced in energy harvesting SCNs (EH-SCNs) network management. In this paper, we study the self- organized cell operation management problem with different quality of service (QoS) requirements of users, in which the EH-SCNs needs to perform cell activation operation in a distributed manner with the uncertainty of harvested energy. With the assumption of Markovian energy harvesting process, multi-armed bandit game (MAB) based Thompson Sampling algorithm is developed to solve the small cell activation problem with a self-organized manner in EH-SCNs. Simulation results show that our proposed approach is particularly suitable to manage the large-scale EH-SCNs more efficiently under uncertain environment with incomplete information.

In this paper, we investigate the multiservice delivery between the source-destination pairs in distributed selfish wireless networks (SeWN), where selfish relay nodes (RN) expose their selfish behaviors, i.e., forwarding or dropping multiservices. Owing to the effect of the RNs' node-selfishness on the multiservices, a distributed framework of the node-selfishness management is constructed to manage the RN's node-selfishness information (NSI) in terms of its available resources, the employed incentive mechanism and the quality-of-service (QoS) requirements, and the other RNs' NSI in terms of their historical behaviors. In this framework, the RNs' NSI includes the degree of node-selfishness (DeNS), the degree of intrinsic selfishness (DeIS) and the degree of extrinsic selfishness (DeES). Under the distributed node-selfishness management, a path selection criterion is designed to select the most reliable and shortest path in terms of RNs' DeISs affected by their available resources, and the optimal incentives are determined by the source to stimulate forwarding multiservices of the RNs in the selected path. Our simulation results demonstrate that this framework effectively manages the RNs' NSI, and the optimal strategies of both the path selection and the incentives are determined.