Search

A novel multi-feature fusion neural network (MFNet) is proposed to address the lack of applicability of existing medical aid diagnostic methods for cross-site and cross-tissue cytopathic lesion screening. MFNet consists of a data-sharing layer, a detailed feature transfer branch, a microscopic identification branch, and an auxiliary loss function. The data-sharing layer converts data images into a feature matrix and extracts detailed elements such as cell morphology, contour, and texture. The microscopic recognition branch obtains multilevel elements by convolving the input elements in stages and fusing them, so that the network can focus on high-level semantic elements such as minor differences of cytopathy. The detail feature transfer branch transfers detail elements across layers and achieves cross-layer fusion with semantic elements. The auxiliary loss function enables the network feature classification capability to be enhanced. MFNet is experimentally compared with AlexNet, VGG-16, ResNet-50, and other models on the tumor cell datasets, and the results show that the proposed method can effectively improve the recognition accuracy.

Melting and evolution of Arctic Ocean sea ice in summer are one of the main factors affecting sea ice albedo and climate. At present, remote sensing images of sea ice are mainly acquired with the help of payload platforms, such as aircraft or satellites. The above methods have a large measurement scale and a short period. Sea ice research also requires suitable equipment for fine-scale, visualized long-term tracking observations of target sea ice regions. This article investigates the feasibility of measuring interannual variability in sea ice morphology in the Arctic sea ice environment using a low-power gimbal camera automated by satellite communications. Year-round observational data from the arctic field were used. The sea ice images were defogged and edge-enhanced by a newly developed algorithm with the help of human visual enhancement principles. During the 11th Chinese National Arctic Science Expedition, the system was placed on perennial ice in the northern Beaufort Basin to monitor sea ice appearance changes and the external state of the ice mass balance buoy. The system fuses visually enhanced sea ice images with ice balance buoy data from the same ice floe in a multi-scale environmental sequence to obtain sea ice observations with image sequences. The data complement the need for long-duration, fixed-ice, and short-visual-range observations in sea-ice studies. The equipment design methods and image processing algorithms involved in the experimental process have accumulated engineering experience for this application area. The research methodology provides a new framework for long-term, and visualized sea ice observation methods.

With the rapid development of 5G technology in recent years, multimedia communication services, such as live online and short video, have received wide attention and become the important means of people's daily social intercourse. However, the rapid growth of multimedia communication demands pose greater challenges to both the wireless network communication capacity and the network processing capacity. Mobile Edge Computing (MEC) is widely regarded as a promising technology to cope with the above challenges. To satisfy the growing demands and improve the quality of experience for users, it is in urge need to seek the effective and efficient task assignment and computing offloading strategy for MEC networks. In this paper, we focused on the multimedia services which need to be processed, uploaded and shared in the network and research the long-term task assignment and resource coordination problem. We formulate the optimization problem as a stochastic optimization problem with the aim of the minimizing the time-average energy consumption of the system. By using the Lyapunov optimization technique, we decompose the original problem into several subproblems which can be solved with current system information and low computational complexities. On this basis, we propose an online energy-efficient task assignment and computing offloading strategy to adaptively decide the task assignment, coordinate and optimize the wireless and computation resource allocation by taking the dynamic wireless condition and service delay constraints into consideration. Extensive simulation results show that our proposed algorithm can achieve considerate energy consumption and delay performances under different conditions.

Precise time synchronization becomes a vital constituent due to the rigorous needs of several time-sensitive applications. The clock synchronization protocol is one of the fundamental factors that can define the quality of communication. Our study starts with a brief discussion on the application domain of precise time synchronization and comes with an in-depth study of the synchronization with the main focus on the IEEE 1588 Precision Time Protocol (PTP). We have compared the well-known synchronization techniques and conclude that the PTP is the most appropriate answer to robust clock synchronization though challenges are there that requires thoughtful efforts and modification in the current version. The working mechanism and main components of the PTP network are discussed. We have established a testbench using commercially available devices and development boards to evaluate the PTP performance under different configurations. Major sources of synchronization error and other aspects contributing to precision are examined. This paper discussed numerous approaches that could enhance the performance of the PTP protocol. Structures for PTP based wireless clock synchronization required by advanced applications has also been discussed. In the end, paper focuses on the main industrial application areas in which PTP plays an important role, including WLAN, optical data centers, Smart grid, IEC 61850, etc. We conclude the paper by identifying the future trends and research directions for PTP based clock synchronization.

Multiply-accumulate calculations using a memristor crossbar array is an important method to realize neuromorphic computing. However, the memristor array fabrication technology is still immature, and it is difficult to fabricate large-scale arrays with high-yield, which restricts the development of memristor-based neuromorphic computing technology. Therefore, cascading small-scale arrays to achieve the neuromorphic computational ability that can be achieved by large-scale arrays, which is of great significance for promoting the application of memristor-based neuromorphic computing. To address this issue, we present a memristor-based cascaded framework with some basic computation units, several neural network processing units can be cascaded by this means to improve the processing capability of the dataset. Besides, we introduce a split method to reduce the pressure of the input terminal. Compared with VGGNet and GoogLeNet, the proposed cascaded framework can achieve 93.54% Fashion-MNIST accuracy and 86.64% CIFAR-10 accuracy under the 4.15M parameters. The extensive experiments with Ti/AlOx/TaOx/Pt we fabricated are conducted to show that the circuit simulation results can still provide a high recognition accuracy, and the recognition accuracy loss after circuit simulation can be controlled at around 0.39%.

Many mainstream applications require multiply-accumulate calculations, such as image processing and neuromorphic computing. Multiply-accumulate calculations using memristor crossbar arrays is a remarkable method for extremely high implementation efficiency, whereas the memristor array fabrication technology is still not mature and it is difficult to fabricate large-scale arrays with high-yield, which will seriously affect the performance of the application running on the RRAM crossbar. This paper proposes an inputs split based calibration method that improves the application accuracy in tolerating variations and stuck-at-fault of memristor devices. To demonstrate the performance of the calibration algorithm, the case of image sharpening and three neural networks architectures are applied for simulation experiments. And experimental results show that the calculation accuracy can be improved by up to 26.83% at 90% yield of crossbar arrays, and the success rate of the algorithm can be as high as 99.3% when there are several arrays cascaded. It is of great significance to the application of arrays in multiply-accumulate computation.

Various types of knowledge and features have been explored for level set-based segmentation. On the ground, the prior knowledge and carefully-designed features perform well to identify the foreground-background contrast, which improves the performance of the segmentation method for complicated and distorted data. However, this is not the case for underwater environments, since the features available on the ground are not suitable for challenging underwater environments. Thus, underwater image segmentation currently lags behind ground-based segmentation. In this paper, novel cues and a suitable model formulation for object segmentation from underwater images are proposed. We consider the special haze effect over underwater images and extract an informative feature (transmission feature) from haze condensation. The saliency feature is also used for underwater object segmentation. Consequently, in our method, the object-background difference can be presented by these features on two levels, i.e., the edge-level transmission and region-level saliency features. These two types of features are integrated into a unified level set formulation to propose a solution that handles the challenging issues in underwater object segmentation. The experimental comparisons of our method with other methods comprehensively demonstrate the satisfactory performance of our method.

With the rapid development of mobile communication technology, short video applications, which combine the features of both social and multimedia applications, have become more and more popular. However, the transmission of short videos poses great challenges to the existing mobile networks. In this paper, mobile edge computing is adopted to provide content caching of short videos close to end users. To improve the quality of service, we take both the quality level of the video and the long-term wireless network transmission performance into consideration. The joint video quality selection and radio bearer control optimization problem is formulated as a Markov decision process, aiming at maximizing the long-term video quality profit and minimizing the cost of bearers and the penalty of latency. Deep reinforcement learning is used as the solution and the policy gradient based quality selection and radio bearer control method is proposed. The REINFORCE algorithm with baseline is used to train the policy network, and the episodic simulations are built to obtain the training samples. Different weight coefficients of the objective function are configured. Training results show that the proposed method can achieve the best accumulated value among all the comparison methods. When the weight coefficients are changed, the training processes can lead the policy networks to obtain proper trade-off between different objective factors. Moreover, the performance of the trained policy network is evaluated with different short video request arriving rates. Testing results show that the proposed method performs well when the arriving rates vary in a certain range.

The Hermiticity of physical observable is the fundamental axioms in quantum mechanics. However, it is found that several non-Hermitian Hamiltonian still have real eigenvalues, especially the parity-time symmetric system. The property of parity-time symmetry leads the system to various nontrivial physics with interesting counterintuitive features around the exceptional point. In this study, we investigate the dynamics of parity-time-anyonic Hamiltonian under the tunable exceptional points related to an arbitrary phase of the system. We show that the parity-time symmetry can be achieved in a quasi-parity-time-symmetric system. Also the phase dependence symmetry is discussed based on the simulated results.

As the continuous development of mobile social networks, the structure of the mobile social network increasingly becomes complex. It not only speeds up information transmission between people but also expands the scope of information exchange, which has become an essential and important social media in people's social life. How to effectively identify and classify these online communities has important practical significance for the study of social networks. Correctly detecting the community structure of mobile social networks can not only improve the accuracy of friend recommendation, link prediction, service user positioning, product marketing, and other aspects but also provide an important basis for the monitoring of online public opinion. But the traditional social network cluster method based on the trust degree mainly calculates the user trust by analyzing the interactive feedback information between users. This method cannot effectively solve the “cold start” problem in the trust calculation process, that is, for the new network node, the trust value of this node cannot be accurately measured due to the lack of interaction with other nodes. Focusing on this problem, we propose a Gaussian pigeon-oriented graph clustering algorithm for social networks' cluster in this paper. A graph model is first built. Then, an efficient K-medoid algorithm is utilized to search the user center in all groups. The Gaussian pigeon algorithm is used to search the similarity between each user and the central user. Users that meet the similarity threshold are divided into the same user group. Finally, the simulation results show that the proposed method has better cluster effect than other state-of-the-art social networks' clustering approaches.

The robust and efficient detection of infrared small target is a key technique for infrared search and track systems. Several robust principal component analysis (RPCA)-based methods have been developed recently, which have achieved the state-of-the-art performance. However, there are still two drawbacks: 1) the false alarm ratio would raise under the heavy background clutters and noises and 2) these methods are usually time-consuming and not suitable for real-time processing. To solve this problem, an infrared patch-tensor (IPT) model based on weighted tensor nuclear norm is proposed in this paper. First, the infrared image is transformed into the IPT. Considering the sum of nuclear norms adopted in the IPT model is not the convex envelope of the tensor rank, and the solution is substantially suboptimal. The tensor nuclear norm is adopted to recover the underlying low-rank background tensor and sparse target tensor, and the computation complexity can be reduced dramatically with the help of the tensor Singular Value Decomposition. Moreover, to further suppress the background clutters, a weight tensor is incorporated with tensor nuclear norm to preserve the background edges better. Then, the separation between target and background is formulated as a convex weighted tensor RPCA model. Finally, the proposed model can be solved by the alternating direction method of multipliers. Extensive experiments demonstrate that the proposed model outperforms the other state-of-the-arts in terms of performance and efficiency.

In recent years, a new concept of network architecture with control/data plane separation is introduced to the future network, which can effectively improve the mobility robustness and reduce the handover failure by providing the data plane services under the umbrella of a macro cell coverage layer. However, in the ultra dense network, frequent data plane handovers would still introduce huge signaling exchanges and latencies. In this paper, we propose a universal predictive mobility management scheme for urban ultra-dense networks to speed up the data plane handover process. We utilize the probability suffix tree model to save and analyze the transition relationships between small cells in terms of variable markov chains, and pre-configure a cluster of small cells with larger handover probabilities for the users. To accommodate different versions of the users, a compatible network-controlled predictive mobility management procedure and an advanced user-autonomous predictive mobility management procedure are proposed to support the proposed predictive mobility management scheme. The simulation results show that the proposed scheme can significantly improve the prediction accuracy with a lower redundant configuration cost and can effectively speed up the data plane handover process compared with the traditional mobility management.