Speckle noise removal based on structural convolutional neural networks with feature fusion for medical image

Application of convolutional neural networks (CNNs) for image additive white Gaussian noise (AWGN) removal has attracted considerable attentions with the rapid development of deep learning in recent years. However, the work of image multiplicative speckle noise removal is rarely done. Moreover, most of the existing speckle noise removal algorithms are based on traditional methods with human priori knowledge, which means that the parameters of the algorithms need to be set manually. Nowadays, deep learning methods show clear advantages on image feature extraction. Multiplicative speckle noise is very common in real life images, especially in medical images. In this paper, a novel neural network structure is proposed to recover noisy images with speckle noise. Our proposed method mainly consists of three subnetworks. One network is rough clean image estimate subnetwork. Another is subnetwork of noise estimation. The last one is an information fusion network based on U-Net and several convolutional layers. Different from the existing speckle denoising model based on the statistics of images, the proposed network model can handle speckle denoising of different noise levels with an end-to-end trainable model. Extensive experimental results on several test datasets clearly demonstrate the superior performance of our proposed network over state-of-the-arts in terms of quantitative metrics and visual quality.     

Reliability evaluation for distributed computing networks withimperfect nodes

This paper examines two compensating methods that: (1) account for imperfect nodes, and (2) can be embedded in most symbolic network reliability algorithms that presume perfect nodes. The Aggarwal method can be exponential in time with the number of links, whereas the Torrieri method is always linear. However the Torrieri method can yield incorrect results for some undirected networks. This paper points out such incorrectness and then proposes an efficient reliability evaluation algorithm (ENR/KW) accounting for imperfect nodes in distributed computing networks. Based on the concept of network partition, ENR/KW exploits some simple efficient techniques to handle the unreliable nodes, for directly computing the network reliability expression considering imperfect nodes instead of using any compensating method. The basic idea of ENR/KW is to partition the network directly into a set of smaller disjoint subnetworks by only considering link elements as if all nodes are perfect. Each disjoint subnetwork is generated by maintaining a specific directed graph structure to consider the effect of imperfect nodes. Therefore, the reliability expression for imperfect nodes can be obtained directly from the disjoint subnetwork and the specific directed graph. ENR/KW can be generalized to evaluate various network reliability measures considering imperfect nodes such as terminal-pair reliability, K-terminal reliability, and distributed-program reliability. Many experiments for evaluating the terminal-pair reliability and distributed-program reliability were performed on a SUN workstation to show the efficiency of ENR/KW in terms of the number of generated subnetworks and overall computation time     

Intrusion Detection Techniques for Mobile Wireless Networks

The rapid proliferation of wireless networks and mobile computing applications has changed the landscape of network security. The traditional way of protecting networks with firewalls and encryption software is no longer sufficient and effective. We need to search for new architecture and mechanisms to protect the wireless networks and mobile computing application. In this paper, we examine the vulnerabilities of wireless networks and argue that we must include intrusion detection in the security architecture for mobile computing environment. We have developed such an architecture and evaluated a key mechanism in this architecture, anomaly detection for mobile ad-hoc network, through simulation experiments.     

Intelligent control of complex electrochemical systems with aneuro-fuzzy-genetic approach

This paper describes different hybrid approaches for controlling the battery charging process. The hybrid approaches combine soft computing techniques to achieve the goal of controlling the temperature of the battery during the electrochemical charging process. We have reduced the time required for charging a battery with the use of fuzzy logic, neural networks, and genetic algorithms. In the neuro-fuzzy-genetic approach, neural networks are used for modeling the electrochemical process, fuzzy logic is used for controlling the process, and genetic algorithms are used to optimize the fuzzy system     

Cascade neural networks for multichannel blind deconvolution

The authors present a new simple but efficient and powerful extension of Bussgang-type blind equalisation algorithms which can extract multiple source signals from their unknown convolutive mixtures. A cascade neural network is proposed, where each module consists of an equalisation subnetwork and a deflation subnetwork. This approach can adopt any blind equalisation algorithm (which has been developed for the equalisation of a single channel). It can also be applied when the number of source signals is not known in advance. Extensive computer simulation results confirm the validity and high efficiency of the proposed method     

Design of Translucent Optical Networks: Partitioning and Restoration

We discuss the problem of designing translucent optical networks composed of restorable, transparent subnetworks interconnected via transponders. We develop an integer linear programming (ILP) formulation for partitioning an optical network topology into subnetworks, where the subnetworks are determined subject to the constraints that each subnetwork satisfies size limitations, and it is two-connected. A greedy heuristic partitioning algorithm is proposed for planar network topologies. We use section restoration for translucent networks where failed connections are rerouted within the subnetwork which contains the failed link. The network design problem of determining working and restoration capacities with section restoration is formulated as an ILP problem. Numerical results show that fiber costs with section restoration are close to those with path restoration for mesh topologies used in this study. It is also shown that the number of transponders with the translucent network architecture is substantially reduced compared to opaque networks.     

平坦瑞利环境下多用户检测的量子神经网络方法

提出采用量子神经网络(QNN)方法在平坦瑞利环境下进行多用户检测的方法。量子神经网络是量子计算与人工神经网络(ANN)相结合的产物,由于利用量子并行计算和量子纠缠等特性从而克服了传统人工神经网络的固有缺点。研究结果表明:该算法具有较强的鲁棒性;能有效地抑制噪声干扰,克服远近效应,在平坦瑞利衰减下具有较好地误码性能。     

The use of artificial neural networks for optimal message routing

Artificial neural networks are being designed to exploit the unique computational power of the human brain. A brief review of neural networks and their applications in communications is presented. Following that, the neural network solutions to the routing problems are given. Simulation results and performance comparisons are discussed. A comparison between the neural approach and other popular routing algorithms such as Bellman-Ford's and Dijkstra's algorithms is then presented. The practical significance of this new routing algorithm is discussed and further research work is suggested     

子网络抽取定理和多端反馈网络的拓扑公式

本文提出了混合图关于二点对和超边分解的变形图的概念,应用它们和有向超图理论导出了参数抽取定理和子网络抽取定理的拓扑公式,进而导出了多端反馈有源网络的拓扑公式。公式中反馈子网络与基本子网络的参数是分开的,便于看出反馈参数的影响;而且由于把一个网络分解成二个较小的子网络,可以降低计算的时间复杂度和空间复杂度。     

Winner-take-all neural network and its application to minimax optimisation problems

A new and simple winner-take-all neural subnetwork suitable for VLSI CMOS implementation is proposed. The subnetwork is used for the real-time solving of minimax optimisation problems. In particular, the Letter shows how to solve, in real time, an over-determined system of linear equations by using the Chebyshev L/sub infinity / norm criterion and a neural network approach. The validity and performance of the network have been checked by extensive computer simulation of different minimax optimisation problems.<>     

TOPOLOGICAL FORMULAS FOR SUBNETWORK EXTRACTION THEOREMS AND MULTITERMINAL FEEDBACK NETWORKS

The concepts of modified graphs of a composite graph with respect to two vertex-pairs and a hyperedge-decomposition are introduced,respectively.By applying them and thedirected hypergraph theory,the topological formulas for the parameter-extraction theorem andsubnetwork-extraction theorems are derived,and then the topological formulas for multiterminalfeedback networks are presented.In these formulas the parameters of the feedback subnetworkare separated from that of the fundamental subnetwork,so that it is convenient to find out theeffect of the feedback parameters.Furthermore,since one network is decomposed into two smallersubnetworks,the computing time complexity and space complexity can be reduced.     

基于模糊神经网络的网络业务量预测研究

利用神经网络(NN)的自学习能力以及模糊逻辑的动态性和及时性等特点,将模糊逻辑和 NN 有机地结合起来,构造出了五层模糊神经网络(FNN),并用训练 NN 的相应学习算法-BP 算法来训练网络。本文将 FNN 用于网络自相似业务预测研究中,并与单纯的 NN 算法相比较。仿真结果表明,FNN 能很好地预测复杂网络业务,与传统的 NN 算法相比,不仅收敛速度快,且得到更好的预测效果。本文为复杂网络业务流量预测研究提供了一种有效途径。     

A minimum-energy path-preserving topology-control algorithm

The topology of a wireless multihop network can be controlled by varying the transmission power at each node. It is not energy efficient to use the communication network G/sub max/ where every node transmits with maximum power. For energy efficient operations, it is desirable to have a subnetwork that preserves a minimum-energy path between every pair of nodes (where a minimum-energy path is one that allows messages to be transmitted with a minimum use of energy). We first identify conditions that are necessary and sufficient for a subnetwork G of G/sub max/ to preserve this property. Using this characterization, we then propose an efficient topology-control algorithm that, given a communication network G/sub max/, computes a subnetwork G that it preserves at least one minimum-energy path between every pair of nodes. We also propose an energy-efficient reconfiguration protocol that maintains this minimum-energy path property as the network topology changes dynamically. We demonstrate the performance improvements of our algorithm over other existing topology-control algorithms through simulation.     

VLSI hardware for example-based learning

Example-based learning, as performed by neural networks and other approximation and classification techniques, is both computationally intensive and I/O intensive, typically Involving the optimization of hundreds or thousands of parameters during repeated network evaluations over a database of example vectors. Although there Is currently no dominant approach or technique among the various neural networks and learning algorithms, the basic functionality of most neural networks can be conceptually realized as a multidimensional look-up table. While multidimensional look-up tables are clearly impractical due to the exponential memory requirements, we are pursuing an approach using interpolation based only on the sparse data provided by an initial example database. In particular, we have designed prototype VLSI components for searching multidimensional example databases for the X closest examples to an input query as determined by a programmable metric using a massively parallel search. This nearest-neighbor approach can be used directly for classification, or in conjunction with any number of neural network algorithms that exploit local fitting. The hardware removes the I/O bottleneck from the learning task by supplying a reduced set of examples for localized training or classification. Though nearest-neighbor retrieval algorithms have efficient software implementations for low-dimensional databases, exhaustive searching is the only effective approach for handling high-dimensional data. The parallel VLSI hardware we have designed can accelerate the exhaustive search by three orders of magnitude. We believe this special purpose VLSI will have direct application in systems requiring learning functionality and in accelerating learning applications on large, high-dimensional databases     

Sequential capacity determination of subnetworks in network performance analysis

In the process of performance evaluation for a stochastic network whose links are subject to failure, subnetworks are repeatedly generated to reflect various states of the network, and the capacity of each subnetwork is to be determined upon generation. The capacity of a network is the maximum amount of flow which can be transmitted through the network. Although there are existing algorithms for network capacity computation, it would create a great number of repetitions to compute the capacity of each subnetwork anew upon generation in the process. This is true because subnetworks are generated by combining certain links to the current one, and hence each current subnetwork is embedded in those new subnetworks. Recently, a number of methods have been proposed in the context of searching a method which efficiently computes the capacity of subnetworks by utilizing the given information of minimal paths, and preferably without many repetitions in sequential computations. But, most of the methods still have drawbacks of either failing to give correct results in certain situations, or computing the capacity of each subnetwork anew whenever the subnetwork is generated. In this paper, we propose a method based on the concepts of signed minimal path, and unilateral link, as defined in the text. Our method not only computes the capacity of each subnetwork correctly, but also eliminates the repetitive steps in sequential computations, and thereby efficiently reduces the number of subnetworks to consider for capacity computation as well. Numerical examples are presented to illustrate the method. The drawbacks of other methods are also discussed with counter examples.     

Routing in optical multistage interconnection networks: a neuralnetwork solution

There has been much interest in using optics to implement computer interconnection networks. However, there has been little discussion of any renting methodologies besides those already used in electronics. In this paper, a neural network routing methodology is proposed that can generate control bits for a broad range of optical multistage interconnection networks (OMIN's). Though we present no optical implementation of this methodology, we illustrate its control for an optical interconnection network. These OMIN's can be used as communication media for distributed computing systems. The routing methodology makes use of an artificial neural network (ANN) that functions as a parallel computer for generating the routes. The neural network routing scheme can be applied to electrical as well as optical interconnection networks. However, since the ANN can be implemented using optics, this routing approach is especially appealing for an optical computing environment. Although the ANN does not always generate the best solution, the parallel nature of the ANN computation may make this routing scheme faster than conventional routing approaches, especially for OMIN's that have an irregular structure. Furthermore, the ANN router is fault-tolerant. Results are shown for generating routes in a 16×16, 3-stage OMIN     

约束优化神经网络建模和控制策略研究

根据人工神经网络的基本优化机理，研究了基于Ｌａｇｒａｎｇｅ函数的适合于求解一般约束问题的神经网络建模方法，探讨了神经元非线性度和拉氏乘子等提高网络优化设计效率的控制策略，测试结果证明了提出的网络和控制策略的可行性和有效性。     

Neural networks for vector quantization of speech and images

Using neural networks for vector quantization (VQ) is described. The authors show how a collection of neural units can be used efficiently for VQ encoding, with the units performing the bulk of the computation in parallel, and describe two unsupervised neural network learning algorithms for training the vector quantizer. A powerful feature of the new training algorithms is that the VQ codewords are determined in an adaptive manner, compared to the popular LBG training algorithm, which requires that all the training data be processed in a batch mode. The neural network approach allows for the possibility of training the vector quantizer online, thus adapting to the changing statistics of the input data. The authors compare the neural network VQ algorithms to the LBG algorithm for encoding a large database of speech signals and for encoding images     

Fault diagnosis of analog circuits

In this paper, various fault location techniques in analog networks are described and compared. The emphasis is on the more recent developments in the subject. Four main approaches for fault location are addressed, examined, and illustrated using simple network examples. In particular, we consider the fault dictionary approach, the parameter identification approach, the fault verification approach, and the approximation approach. Theory and algorithms that are associated with these approaches are reviewed and problems of their practical application are identified. Associated with the fault dictionary approach we consider fault dictionary construction techniques, methods of optimum measurement selection, different fault isolation criteria, and efficient fault simulation techniques. Parameter identification techniques that either utilize linear or nonlinear systems of equations to identify all network elements are examined very thoroughly. Under fault verification techniques we discuss node-fault diagnosis, branch-fault diagnosis, subnetwork testability conditions as well as combinatorial techniques, the failure bound technique, and the network decomposition technique. For the approximation approach we consider probabilistic methods and optimization-based methods. The artificial intelligence technique and the different measures of testability are also considered. The main features of the techniques considered are summarized in a comparative table. An extensive, but not exhaustive, bibliography is provided.     

解无约束极大极小问题的非对称神经网络算法

本文构造了一种新的非对称神经网络模型，用于求解极大极小无约束优化问题，并证明了非对称线性神经网络和非线性神经网络是整体Ｌｙａｐｕｎｏｖ稳定的，且收敛于对应的Ｌａｇｒａｎｇｅ方程的稳定点，计算机模拟的结果表明此方法是可行的，且具有良好的整体收敛性。