Search

This paper proposes a novel relay algorithm to optimize the communication throughput of unmanned aerial vehicle (UAV) mobile relay formations while considering the challenges posed by obstacle avoidance, channel complexity, high dynamics of UAVs, and real-time mission requirements. To tackle the non-convex nature of this problem, we develop the unscented Kalman filter and hybrid particle swarm optimization (UKF-HPSO) algorithm. Initially, real-time prediction of the source and destination of UAV positions is accomplished using the UKF. Subsequently, these predicted coordinates serve as inputs for achieving the optimal deployment of relay UAVs under the constraints imposed by HPSO. The superiority of the UKF-HPSO algorithm compared to baseline approaches is demonstrated through extensive simulations. System throughput is effectively optimized while maintaining real-time performance by our proposed algorithm, which addresses the unique challenges of UAV communication in dynamic environments.

Quality of Service (QoS) prediction for Web services is a hot research topic in the field of services computing. Recently, representation learning of heterogeneous networks has attracted much attention, and specifically the relationship between users and services, as a typical heterogeneous network in which heterogeneity and rich semantic information provide a new perspective for QoS prediction. This paper proposes a novel QoS Prediction scheme based on a heterogeneous graph attention network. Our method first unitizes the user’s location information to construct an attributed user-service network. Then, considering the difference between nodes and links in the latter network, we model a heterogeneous graph neural network based on a hierarchical attention machine (HGN2HIA) that includes node- and semantic-level attentions. Specifically, node-level attention aims to learn the importance between a node and its meta-path-based neighbors, while semantic-level attention learns the importance of different meta-paths. Finally, user embedding will be generated by aggregating features from meta-path-based neighbors in a hierarchical manner, used for QoS prediction. Experimental results on the public WS-Dream dataset demonstrate the superior performance of the proposed model over the current state-of-the-art methods, with NMAE and RMSE metrics reduced by at least 2.56% and 1.3%, respectively. Furthermore, the experimental results highlight that node-level attention contributes more than semantic-level. Overall, we demonstrate that introducing these attention levels improves the QoS prediction performance.

Social Internet of thing (SIoT) is an emerging area that introduces sociology theories into IoT networks, in order to support the spread behaviour of multimedia contents. With the explosive growth of data in the internet, traditional solutions are no longer suitable for large-scale SIoT networks. To address this challenge, this work target to optimize the information propagation in SIoT networks. First, we establish a novel model to divide SIoT network into communities on the basis of network social attributes. Second, we propose an information dissemination control mechanism based on variation trends of content popularity, in order to maximize the spread performance of information. Third, we check the relations between performance indicators and network attributes via a computer simulation. Simulation results show the feasibility of the proposed mechanism. Simulations results also reveal that the proposed mechanism make better balance compared with existing solutions in terms of efficiency and complexity.

Potential risk of X-ray radiation from computed tomography (CT) has been a concern of the public. However, simply decreasing the dose will degrade quality of the CT images and compromise diagnostic performance. In this paper, we propose a noise learning generative adversarial network coupling with least squares, structural similarity and L1 losses for low-dose CT denoising. In our method, noise distributed in the input low-dose CT image is learned by the generator network and then subtracted from the input to generate the final denoised version. The denoised CT images are penalized by the least squares loss function, and they are pulled toward boundary of the decision even though they are classified as normal-dose CT. Least squares stabilize the training process without regularization. Structural similarity and L1 losses are utilized to keep textural details and sharpness of the denoised CT images respectively. Experiments and results show that our method can effectively suppress noise and remove artifacts compared with the state-of-the-art methods. The texture statistical properties, which include mean, standard deviation, uniformity, and entropy, further confirm that the generated noise-reduced CT image is as close as to that of the normal-dose counterpart.

Polychromatic source X-ray fluorescence computed tomography (XFCT) is a novel imaging method, which is applied for the diagnosis and treatment studies of early stage cancer. In this paper, we propose a fast multi-pinhole L-shell XFCT imaging system to reduce the scan time and radiation dose, which has potential in detecting lower concentration of contrast agents (gold nanoparticles, GNPs). Two kinds of phantoms - the concentration-phantom 1 and the size-phantom 2 are designed to verify imaging performance for low-concentration and small-size ROI. The scanning processes are simulated by Geant4, and images are reconstructed by Optimized EM-TV algorithm. It is concluded that this imaging system is more sensitive in detecting low concentration GNPs than K-shell imaging system. Simulation experiments show the reconstructed images can achieve the highest CNR both for phantom 1 and phantom 2 with iterating 10 times. The detection limit can reach 0.16% when pinhole radius is 0.08mm.

As the emergence of numerous services with similar functions, it is very helpful to recommend personalized services for users, and urgent to accurately predict the QoS(Quality-of-Service) values of Web services. Collaborative Filtering (CF) is a commonly-used method to handle above issues. However, it faces two common issues: data sparsity problem and trustworthiness issue, which greatly reduces its prediction accuracy. To address this problem properly and systematically, we introduce the network embedding learning into the QoS prediction process and propose an improved QoS prediction method based on the reputation-aware network embedding learning. Firstly, a two-phase K-means clustering is adopted to filter untrustworthy users. Next, the reputation of trustworthy users is calculated, and an attributed user-service bipartite network is constructed between trustworthy users and services while considering the user reputation. Then the reputation-aware network embedding is adopted to learn the hidden representations of users. Finally, user-based CF is adopted to predict the unknown QoS values. The experimental results show that our method has a significant improvement in accuracy compared with other methods.

In the era of Internet and big data, an increasing number of intelligent applications have been developed. As the result, a lot of user data can be collected and stored by Internet companies as well as by ordinary users through various media platforms such as Facebook, WeChat, etc. that may contain information related to personal privacy. Even though privacy protection has been declared by Internet service providers, after collecting enough amount of seemingly less relevant data, an attacker can still infer user privacy via one means or another, e.g., by running a data mining algorithm. This can undoubtedly bring high risk of privacy disclosure to users under such an attack model. So, accurately measuring the leakage of privacy becomes an urgent issue. Although many privacy measurement and protection methods have been proposed in recent years, they mainly target at structured datasets and are thus inadequate to the measurement of the disclosure of specific privacy information. In addition, most of the methods have failed to consider the internal connections and relationships between privacy information and thus cannot be used to measure the implicit privacy disclosure risk on unstructured data. In this paper, we propose a semantic inference method based on the WordNet ontology to measure privacy disclosure in which we employ an information content (IC) based method to determine the weight of attributes to describe the inference preferences in the process of inferring privacy. Experiment was performed to verify the effectiveness of the IC based inference weight assignment method and to compare the proposed measurement method to some privacy disclosure behavior learned through a data mining algorithm and some existing privacy measurement methods to demonstrate the advantages of the proposed method for measuring privacy disclosure.

The spectral computed tomography (CT) based on photon counting detectors can collect the incident photons with different energy ranges. However, due to the low photon counts in narrow energy bin and the unhomogeneous response problem of detector cells, there are severe noise and ring artifacts in reconstructed spectral CT images. We proposed an image denoising and ring artifacts removal method via improved Fully Convolutional Pyramid Residual Network (FCPRN). In our study, we scanned a mouse specimen with spectral CT based on photon counting detector, and reconstructed mouse CT images as data set. Then we use the data set to train our network for image denoising and ring artifacts removal. Experimental results demonstrated that the proposed method could reduce noise and suppress ring artifacts of spectral CT images concurrently in different energy ranges. And the performance of the FCPRN is better than that of some networks for CT image denoising.

Semantic segmentation networks focus on the scene parsing of an unrestricted open scene. The typical segmentation architectures are stacks consisting of convolutional layers, which are used to extract semantic features. The feature map dimension is sharply changed at sampling units for most of networks, which ensure effective propagation of the gradient in deep nets. In this article, we proposed a state-of-the-art network model named Fully Convolutional Pyramidal Networks (FC-PRNet), which employs pyramidal residual structure to change the feature map dimension at all convolutional layers. This design is an effective way of improving generalization ability and optimizing parameters, and FC-PRNet could achieve excellent capability of semantic extraction. We used urban scene benchmark CamVid and KITTI dataset to test our network, the experimental results show that FC-PRNet achieves better results without any pre-training or post-treatment module. Moreover, due to smart construction of pyramidal residual structures, FC-PRNet has less parameters than other existing networks trained on these datasets.

Network representation learning has attracted much attention as a new learning paradigm to embed network vertices into a low-dimensional vector space, by preserving network information. In this paper, in the light of user co-tag network and social network, we introduced network representation learning techniques into the learning of user preference, to encode user social relations into a continuous vector space. First, we proposed a hybrid network representation learning approach to effectively utilize users' tagging and social relationships, and then we took it for service recommendation. The experimental results show that, compared with four baselines on two public data sets, the improvement ratio over the baselines is up to 50% in terms of Recall@10 and Precision@10 and the improvement is even more than 90% in terms of NDGG@10 and MRR@10.

Screen content video (SCV) is generated by computers, including animation, texts and graphics. SCV contains continuous static frames and many scene changes, making SCV different from conventional camera captured video (CCV) in terms of temporal characteristic. Therefore, conventional quantization parameter (QP) cascading method may not be efficient for SCV. In this paper, a distortion propagation based QP cascading method is proposed for SCV. The special temporal characteristic of SCV is considered and the distortion propagation of every coding tree unit (CTU) is measured. Based on the CTU level distortion propagation, the improved QP cascading method is designed. Experimental results show that compared with other methods, the proposed could achieve better rate distortion (RD) performance and less encoding time.

Molecular communications provide an opportunity to understand the complex mechanisms in living individuals. In this work, we employ the calcium signaling among a group of mobile cells to study the multi-modality behavior of mobile molecular communication networks. We study such behavior from a multimedia view by using the experience of traditional communication networks. In the beginning, we employ a random walk model to study the stochastic mobility of biological cells and dynamic property of cellular networks. Then, we propose a mathematical model to illustrate how calcium signals are generated and propagate in mobile networks. In addition, we propose a multimedia theme to study molecular communications. In this regard, we discuss two transmission forms of calcium signaling, namely concentration transmission and waveform transmission. Further, a simulation is executed to illustrate the patterning of calcium signal waveforms and the results indicate the performance is restrained by environmental factors. The results also exhibit the performance differences of two transmission forms and validate the discussed properties of our proposed multimedia theme.

The spectral computed tomography (CT) based on photon-counting detector performs energy-dependent image reconstruction of material attenuation coefficients, allowing for effective medical diagnosis and material discrimination. However, the spectral CT image quality is degraded in narrow energy bins as a consequence of low photon counts. Thus the edge information of some materials with similar attenuation cannot be well identified. To improve the accuracy of material discrimination of spectral CT, we proposed a deep-learning-based material discrimination method based on Fully Convolutional Pyramidal Residual Network (FC-PRNet). The FC-PRNet model can predict each pixel of spectral CT images and extract more edge information for different material components. We evaluated our method using mouse spectral CT data set, and experimental results demonstrated that the proposed method could efficiently discriminate different materials compared with traditional method based on post-reconstruction. Moreover, our algorithm has fewer parameters, faster convergent speed and higher accuracy, and achieves better quality of material discrimination than SegNet, FCN-8s and U-Net.