Search

The sufficient utilization of the mega data acquired from operative power units is a promising way leading to high efficient, clean and safe operation of power plant. Hereinto, data filtering is a critical link to the data-driven dynamic model identification aiming at optimizing the process control. The machine learning is an advantageous approach for filtering usable data from mega databases for identification due to its effective statistical learning strategy to the field data with noises, disturbances and coupling quantities. Therefore, a data filtering method of combining Gaussian Naive Bayesian classifier and prediction error method (Gaussian NB-PEM) is proposed in this paper. Firstly, variables associated with identification model are selected by analyzing the characteristics of the process. Secondly, the GaussianNB classifier is used for coarse data filtering by calculating the priori probability of each attribute from training sample set and the probability with all possible values of the known categories for testing sample set. Thirdly, the prediction error method is used for further data filtering based on model fitting. By using the filtered closed-loop data, the dynamic characteristics of the superheated steam temperature is modeled and verified by closed-loop control simulation, showing the validity of the Gaussian NB-PEM data filtering method.

Deadlock prevention policies are used to solve the deadlock problems of FMSs. The theory of regions is recognized as the unique method for obtaining maximally permissive (i.e., optimal) controllers in the existing literature. In this paper, the selective siphons and critical markings method [17] is merged in the new deadlock prevention policy. First of all, reachability graph is still needed. Second, CMTSIs are identified. Further, selective siphons and critical markings method is used to check if all dead/quasi-dead markings of CMTSIs are covered by these selective siphons. Furthermore, choose anyone CMTSI that belongs to a same selective siphon to be processed. Finally, controllers are therefore obtained. Experimental results indicate that the computational cost can be reduced again. Besides, it is the most efficient policy to obtain maximal permissive behavior of Petri net models.

The classification of color image that contain skin pixel is a challenging process. It is challenging due to certain factor such as brightness, background similar to human skin represent the obstacles thereof. The basic approach of skin detection is color based classification. The human skin image is the composition of RGB color. The data values of skin is calculated by randomly sampling RGB color values of different face images which have different age, race, genders sets. However, its performance has not really been high because of the high overlapped degree between “skin” and “non-skin” pixels. This paper describes the linear norm fuzzy based twin support vector machine (LN-FTSVM)) approach for discriminate skin and non-skin data values of skin dataset and to enhance the skin recognition performance. The concept of fuzzy is resolved the unclassified and overlapped data region problems. If no decision function is positive for a data set, this data set is classified into a class with the large membership value. By computational experiments shows that Experiments result shows that the accuracy is improved by linear norm fuzzy based TSVM (LN-FTSVM)) over the conventional methods.

In this paper, the generalized detector (GD) constructed on the basis of the generalized approach to signal processing in noise is employed by synthetic aperture radar (SAR) system to define a target location. The GD is proposed to solve the common problem of target detection at the very low signal-to-noise-ratio (SNR) when the conventional detection algorithms are not able to do it. Simulation results demonstrate a GD performance superiority in comparison with the search for undefined maritime objects (SUMO) detector one.

Network coding has been applied successfully in peer-to-peer systems to shorten the distribution time, particularly in extreme conditions where peers in the network are clustered and separated by physical links with limited bandwidth. In this paper, we focus more closely on the use of source coding, i.e. encoding is done only at the source, as an alternative to network coding to facilitate content distribution under such limited bandwidth configuration. We observe analytically and experimentally that, in this specific case, with appropriately chosen expansion factors, source coding can have comparable performance to network coding in terms of distribution time, yet consuming much less computational resources than the latter approach does.

Ontologies have been designed for the purpose of representing a domain of knowledge in a way that makes it understandable by machines and humans. Ontologies have been used in many areas in computer sciences such as knowledge engineering, semantic web, and information retrieval. Religious studies is one of the areas were ontologies are in demand. In this paper, we present the design of domain ontology about prophets and messengers in Islam from trustworthy resources: the Quran and Hadith. The ontology can be used to better understanding all life aspects of prophets and messengers in Islam. The classes, relations and properties of this special domain ontology were gathered and integrated to help researchers to identify and explore knowledge about prophets in Islam such as their lives, books and teachings, the nature of the messages they brought to their nations and tribes and much more.

In this paper, a new configuration for a multi-terminal offshore wind farms based VSC-HVDC and large scale photovoltaic system using modified unified adaptive linear neuron (Adaline) algorithm is presented. In the offshore station, the permanent magnet synchronous generator (PMSG)-based wind turbines are tied to the bus via back-to-back converters and a step-up transformer. The large scale photovoltaic power is connected to the dc bus between the offshore voltage source converters (VSC) and onshore VSC-HVDC stations via a modular multilevel converter (MMC). The Power management and the synchronization between two buses are achieved by controlling the two VSC converters. The performance of the topology proposed is tested using Matlab/Simulink during permanent operation mode and in the presence of DC fault.

The learning behavior of the agent is a challenging and interesting issue in an unknown environment. Reinforcement learning obtain the developed strategy through exploration and interaction with the environment, and the characteristic of online learning make it as an important branch of machine learning research. In this paper, we summarize the current research of the reinforcement learning applications and hotspots. Firstly, the principle, structure and the main classic algorithms of the reinforcement learning are introduced. Secondly, according to the recent research results, we introduce four main applications of reinforcement learning, namely robot, unmanned aerial vehicle, multi-agent and intelligent traffic. Finally, the research hotspots and the development direction of the reinforcement learning are introduced, which conclude the partial perception, hierarchical reinforcement learning, combination with other artificial intelligence technologies and game theory.

Detecting roadside safety attributes and distances in point cloud data is a challenging task. The major problems are accurate detection of attributes and attribute centers for calculating safety distance among attributes. In this paper, we propose a point cloud classification framework for roadside safety attributes detection. In addition, we propose an object center approximation technique for distance calculation that has been integrated into the proposed framework. The proposed framework has been evaluated on large real-world point cloud data, and the experimental results are promising. The framework achieved 100% object-wise accuracy on detecting poles and trees, while the overall point-wise accuracy on detecting all seven attributes was 86%.

In nature, many species became extinct as they could not adapt quickly enough to their environment. They were simply not fit enough to adapt to more and more challenging circumstances. Similar things happen when algorithms are too static to cope with particular challenges of their "environment", be it the workload, the machine, or the user requirements. In this regard, in this paper we explore the well-researched and fascinating family of adaptive indexing algorithms. Classical adaptive indexes solely adapt the indexedness of the data to the workload. However, we will learn that so far we have overlooked a second higher level of adaptivity, namely the one of the indexing algorithm itself. We will coin this second level of adaptivity meta-adaptivity. Based on a careful experimental analysis, we will develop an adaptive index, which realizes meta-adaptivity by (1) generalizing the way reorganization is performed, (2) reacting to the evolving indexedness and varying reorganization effort, and (3) defusing skewed distributions in the input data. As we will demonstrate, this allows us to emulate the characteristics of a large set of specialized adaptive indexing algorithms. In an extensive experimental study we will show that our meta-adaptive index is extremely fit in a variety of environments and outperforms a large amount of specialized adaptive indexes under various query access patterns and key distributions.

This paper presents a new approach for an efficient decimator architecture for cascaded sigma-delta (/spl Sigma//spl Delta/) modulators. The special aspect of the presented structure is the effort to integrate the recombination logic for cascaded modulators in general and additionally a gain error correction for continuous time (CT) modulators in the first stage of the decimator. The design of an appropriate filter topology is derived, analyzed, and verified by simulations. The proposed structure is compared to a conventional implementation, thereby improving efficiency by decades.

Vehicles may move at high speed in vehicular ad-hoc networks (VANETs) and the network topology changes arbitrarily as the nodes move. Vehicles may not find suitable resources because sharing information could quickly become obsolete or irrelevant resources. This paper proposes an efficient geo-aware Peer-to-Peer (P2P) search mechanism to enable resource discovery and sharing in VANETs. We propose discovery and sharing schemes based on the virtual backbone architecture, and then limit the data exchanging region to reduce unnecessary message transmissions. Because the resource node changes very fast, the coordinator must be set to maintain the local node of geographic information in the network. In this scheme, coordinators are established to store information and manage local resource information. In order to ensure dissemination performance, it must guarantee that vehicles possess enough resource pieces to allow efficient bartering. Furthermore, three kinds of piece selection mechanisms, including Nearest-First, Rarest-First, and In-Order are proposed to download resource pieces from other vehicles and to improve the dissemination efficiency. Experiment results demonstrate that our searching and sharing schemes are efficient and outperforms other schemes in terms of satisfactory ratio and download time.

Survey of research papers is not an easy task for novice researchers, because they are not always good at finding all appropriate keywords for the survey. Moreover, it is not easy for them to understand positions of papers in their research fields instantly, even when they use famous search engines like Google Scholar, it may often take a long time for them to find scholarly literature. On the other hand, many researchers have presented citation visualization techniques for surveying research papers. However, it is still often difficult to observe the complicated relations across multiple research fields or traverse the entire relations in their interest. In this paper, we proposed a visualization technique for citation networks applying topic-based paper clustering. Our technique categorizes papers applying LDA (Latent Dirichlet Allocation), and constructs clustered networks consisting of the papers.

This paper describes a new weighting sub-system for fruit sorting and grading machinery. Based on a fixed point DSP, it controls up to 10 lines at 20 fruit/second with an accuracy of /spl plusmn/1 gram. The system applies an adaptive filtering and a posterior deconvolution of the behaviour of the load sensor using an ARMA model. A final estimation procedure to obtain weight in grams is also described. The best results have been obtained with an ALMS algorithm for the adaptive filter.

In technical study programs on the bachelor level the focus is mainly directed on the scientific education in theory and practice. Publication is usually started during the PhD phase when students write their first scientific papers. During the last years some very successful projects were made at the university of applied sciences Regensburg to motivate bachelor students doing their first own small publications. The results of these projects are presented in this paper and ideas how students can publish are given.

Machine vision based automatic apparatus is a necessary component of an automated production line. In this paper, taking can-end surface defect detection as an example, we presented a visual inspector and proposed a new surface defect detection framework. Firstly, the Region of Interest (ROI) of the input image is got by circle detection and prior radius knowledge of can-end. Then, the saliency map and texture filtering image are computed to overcome the problems of irregular texture and non-uniform illumination. Finally, all suspected defects are obtained by saliency map segmentation, connected component analysis. Following, the size, the average saliency value and the mean texture filtering value of each suspected defective region are calculated, and the 3-dimensional features are used to distinguish true defects in the central panel region of can-end. We employ the designed equipment to grab can-end images and do tests divided into four groups on the grabbed images. The accuracy is 98.16% and the execution time is about 65.7 milliseconds. Experimental results demonstrate that the proposed method can inspect faint and low contrast defect accurately. Meanwhile, the proposed method is simple and computationally efficient and meets the high speed of industrial automation.

Video synopsis provides a means for fast browsing of activities in video. Principal background selection (PBS) is an important step in video synopsis. Existing methods make PBS in an offline way and at a high memory cost. In this paper we propose a novel background selection method, ``online principal background selection'' (OPBS). The OPBS selects n principal backgrounds from N backgrounds in an online fashion with a low memory cost, making it possible to build an efficient online video synopsis system. Another advantage is that, with OPBS, the selected backgrounds are related to not only background changes over time but also video activities. Experimental results demonstrate the advantages of the proposed OPBS.

This paper proposed smart agricultural framework over RCAS in CATV network. The proposed framework adopts RCAS, which is a security system which is dedicated to cable television system, as a data protection system, and cable two-way broadband network as a communication channels among smart agricultural service participants.

Synthetic Aperture Radar (SAR) image classification is becoming increasingly important in military or scientific research. SAR image classification based on unsupervised learning usually requires optimization of some metrics. Local optimization techniques frequently fail because functions of these metrics with respect to transformation parameters are generally nonconvex and irregular and, therefore, global methods are often required. In this paper, a new evolutionary approach, particle swarm optimization, is adapted for SAR image classification. The new algorithm composes of three main processes: firstly, selecting training samples for every region in the SAR image. Secondly, training these samples using PSO, and obtain clustering center of every region. Finally, output the classification result of SAR image according to clustering center obtained. To show the effectiveness of this approach, experiment with simulated SAR image was considered. The classification results are evaluated by comparing with two well-known algorithms, K-means and fuzzy K-menas. According to the overall accuracy and Kappa coefficient, PSO has high classification precision and can be used in SAR images classification.

The massive big data transfer requirements in the cloud age are posing unprecedented challenges on todays' network infrastructure. Delivering data in a cost effective and scalable manner is of critical importance to support the continued growth of cloud services. In this paper, we revisit the problem of delivering large data on top of dynamic circuit switched networks, with a particular focus on how it can be used together with packet switching in an integrated data delivery model. We argue that the Software Defined Networking (SDN) concept and flow based switching are crucial to realize scalable and energy efficient integrated data delivery. We propose to use built-in or public storage to resolve contentions between large data flow transfers.

Design of analog and digital electronic circuits is an integral part of the electrical engineering curriculum. Software tools, which enable design, simulation, measuring and circuit analysis as well as PCB (Printed Circuit Board) design, are called EDA (Electronic Design Automation). The paper describes the introduction of EDA software in the first semester of the professional study of electrical engineering at the Zagreb University of Applied Sciences within course Personal Computers in Electrical Engineering. Examples of laboratory exercises related to electrical circuit diagrams, use of component libraries, work in standard laboratory environment (power supplies, function generators, voltmeters, ammeters, oscilloscopes, etc.), circuit analysis in a DC and AC mode as well as in time and frequency domain are described. Focus is on the optimal implementation of EDA tool NI Multisim and its integration with the existing learning management system Moodle as a platform for exercise learning and knowledge assessment.

Truncated multiplication can be used to significantly reduce the power dissipation for applications that do not require correctly-rounded results. This paper presents an efficient method for truncated multiplication called hybrid-correction truncation that utilizes the advantages of two previous methods to obtain lower average and maximum absolute error. Comparisons are presented contrasting power, area, and delay for all three methods compared to standard parallel multipliers. Estimates indicate that hybrid truncated multipliers dissipate slightly less power and consume slightly less area than previous methods for truncated multiplication. In addition, utilization of the hybrid truncation method can provide a method for altering the implementation within certain limits to meet a given precision.

System losses, power generation and generation cost are some of the most important factors that significantly influence the efficiency of power systems. Integrating distributed generation systems into distribution networks is one of the effective solutions. Many studies focused on finding the optimal planning of distributed generation systems to get the maximum benefits from them. In this paper, a brief recent review on a variety of heuristic optimization techniques applied in literature to optimally allocate and size distributed generations will be presented.

Elasticity makes cloud computing an attractive platform for executing complex large-scale expensive dataflows, as it enables different trade-offs between execution time and monetary cost, by varying the number of resources to be provisioned. With cloud providers offering heterogeneous types of resources with different performance and price characteristics, the problem of identifying the various trade-offs available is a great challenge, as the number of possible alternative configurations increases significantly compared to a homogeneous environment, which is itself already computationally difficult. This paper proposes a novel algorithm for dataflow scheduling on heterogeneous clouds that identifies solutions (schedules) close to the optimal pareto front, by exploring the search space in an efficient way. The results of an experimental comparison with the state of the art show that, in several cases, the proposed algorithm provides a richer, more diverse set of solutions, several of which are characterized by significantly better time-money trade-offs.

This paper presents a new voltage-mode multifunction filter with four inputs and three outputs employing three current controlled differential difference current conveyors (CCDDCCs) and two grounded capacitors. The proposed circuit can provide low-pass, band-pass, high-pass, band-stop and all-pass voltage responses by appropriately connecting the input and output terminals, at a high impedance input terminal which enable easy cascadability. The circuit parameters ω 0 and Q can be set orthogonally by adjusting the circuit components. PSPICE simulation results are also performed to confirm the theoretical analysis.

Wireless sensor networks possess significant limitations in storage, bandwidth, and power. Additionally, real-time sensor networks cannot tolerate high latency. While some good compression algorithms exist specific to sensor networks, in this paper we present an energy-efficient method with high compression ratio that reduces latency, storage and bandwidth usage further in comparison with some other recently proposed algorithms. Our Huffman style compression scheme exploits temporal locality and delta compression to provide better bandwidth utilization in the network, thus reducing latency for real time applications. Our performance evaluations show comparable compression ratios and energy savings with a significant decrease in latency compared to some other existing approaches.

Given an image, scene understanding is the process of segmenting and identifying the objects present, and classifying the overall scene. Several frameworks already exist to perform these tasks coherently but training of their probabilistic models is time consuming thereby limiting their scalability. This paper presents a scalable framework adopting an object-based approach. The steps taken by the algorithm are saliency detection for unsupervised object discovery, graph-cut for object segmentation, bag-of-features for object classification and binary decision trees for scene classification. A region of interest (ROI) detector is proposed to automatically provide object location priors from saliency maps for graph-cut. We tested our system on a novel NUS/NTU dataset and compared the scene classification accuracy using different classifiers. Unlike other existing frameworks, the proposed algorithm is scalable and can easily accommodate more object and scene classes.

Current component interfaces, such as those defined by CORBA, are devoid of additional constraints that make helps in the task of composing software, therefore the dependent relations and interaction protocols among components are hidden within the implementations. The "implicity" reduces the composability and makes it difficult to reuse, validate and manage components. This paper presents a message based architectural model and an XML-message based Architecture Description Language (XADL). XADL promotes the description of dependent relations from the implementation level to the architectural level and enhances the interface specification by adding the sequencing, behavior and quality constraints. Furthermore, we formulate the notion of architectural mismatch and propose a checking algorithm to prevent systems from potential mismatches. The use of XML facilitates not only the specification descriptions but also the system implementations, and increases the openness of the whole system as well.

Robustness, security, anonymity, and scalability are valued features for mobile ad hoc network routing protocols. Unfortunately, the existing ad hoc routing protocols have limitations in each area. In this paper, we first analyze several existing mobile ad hoc routing protocols for robustness, security, anonymity, and scalability. Based on the analysis, we propose a new robust anonymous ad hoc on-demand routing protocol based on a proposed anonymous neighborhood trust model. We demonstrate that the protocol has strong security and anonymity protection, better scalability and is more robust when compared with other protocols., MILCOM 2009 - 2009 IEEE Military Communications Conference, October 18-21 2009, Boston, MA, USA

The paper presents a novel short-range wireless sensor node architecture, based on Integral Pulse Frequency Modulator (IPFM) and Ultra-Wideband (UWB) pulse generator. Due to the lack of internal clock signal source and multi user implementation without application of a microprocessor, the architecture is simple and energy efficient. Multi-user coding is performed using delay elements, where each user has unique delay time value. The output of the IPFM is fed to the delay element. Delayed and original signal from the IPFM are feeding UWB pulse generator. The paper presents the transmitter architecture and measurements of the transmitter signals in time and frequency domain.

The construction and the deployment of the service bearer network contain multiple network nodes. So it needs to do a unified operation. In the deployment processing, the whole service instance will fail if one operation fails. So it is a distributed transaction with ACID properties. This paper has done some research on the distributed transaction in a service bearer network. In addition to the basic concepts of distributed transaction. It has done a deep research on the following three aspects: the parallel mechanism of the subtransactions, the processing of the subtransaction failure and priority deployment strategies for the subtransactions. The key point is the designation of a more effective resolution algorithm for the preparation failure of the subtransactions, and then combined with relative examples to validate the efficiency of the algorithm.

According to the environment adaptability require-ment of an autonomous agent system, this paper studies the mapping relationships between the dynamic environment aspects and task scheduling, and then, gives the definitions and computation models of the task, quality of scheduling (QoS) and fitness. Considering the dynamic and inaccurate parameters of environment, a new Fuzzy-logic based Environment Adaptive Scheduler (FEAS) is proposed. In FEAS, the Multi-State Fuzzy Controller (MSFC) can make decision with the environment aspects to adjust the scheduling quality of the multi task groups so as to ensure that the execution process can suit to dynamic environment as possible. Finally, an experiment to make an agent system with FEAS avoids obstacle autonomously is carried out. And the results prove that such task scheduling method we studied can improve its adaptability of an agent system.

This paper presents an iterative, yet effective minimum mean square error (MMSE) space-time multiuser detector (MUD) with the assistance of soft interference to enhance the performance in the multipath code division multiple access (CDMA) systems. The proposed space-time MUD is of rank one and comprises of a spatial filter and a temporal filter. To offset the performance loss due to such a constraint, the new MUD alternatively updates the spatial filter or the temporal filter with the soft interference cancellation in between so as to refine the received signal. Computer simulations show that the new MUD can drastically enhance the performance compared to previous works with modest computations, especially in highly loaded scenarios.

In this paper we propose a Sliding Mode Controller for a class of scalar bilinear systems with delay in both the input and the state. Such a class is considered since it has shown to be suitable for modelling and control of a class of turbulent flow systems. The stability and robustness analysis for the reaching phase in the controlled system are Lyapunov-based. However, since the sliding dynamics is infinite dimensional and described by an integral equation, we show that the stability and robustness analysis is simplified by using Volterra operator theory.

The high energy consumption has become one bottleneck in the development of the data centers (DCs), where the main energy consumers are the cooling system and the servers. Therefore, the joint optimization for the energy efficiency of the cooling system and the servers is a crucial problem, while most of previous works on energy saving only studies one of these two components in an isolated manner. In this paper, we propose a real-time strategy, rTCS (real-time Task Classification and Scheduling strategy), to jointly optimize the energy efficiency of these two components in the scenario where the tasks arrive dynamically. Strategy rTCS first labels the tasks to classify them according to their run time and end time with a time complexity of O(1) and a bounded space complexity. Then, rTCS schedules the tasks in real time based on their labels and the energy consumption model of the DC. Simulation results show that rTCS can effectively improve the energy efficiency of DCs.

Reliability and fault-tolerance are essential requirements of critical, autonomous computing systems. In this paper, we propose a methodology to quantify, and maximize, the reliability of computation in the presence of transient errors when considering the mapping of real-time tasks on an homogeneous multiprocessor system with voltage and frequency scaling capabilities. As the likelihood of transient errors due to radiation is environment- and component-specific, we use machine learning to estimate the actual fault-rate of the system. Furthermore, we leverage probability theory to define a trade-off between power consumption and fault-tolerance. If a processing element fails, our methodology is able to re-map the application, establishing whether the real-time requirements will still be met, and how reliable the new, impaired system will be. Results show that the proposed methodology is able to adjust mapping and operating frequencies in order to maintain a fixed level of reliability for different fault-rates.

Due to the existing cryptographic schemes based on the satisfiability problem (SAT) lack the specific security model and the security analysis of the provable security theory, the security is questioned. For this problem, combining with the advantages of SAT and Elgamal scheme, we construct a secure, verifiable encryption scheme in this paper. Firstly, combining and improving the SRR(N, k, s) model based on the difficulty problem of the random regular formula near the phase transition point, we construct an encryption algorithm. Secondly, on the basis of the existing Damgârd Elgamal scheme, a more secure scheme based on SAT and variants of discrete logarithm problem is proposed. Lastly, the analysis results show that compared with the existing SAT-based hybrid cryptographic schemes, the private key of the hybrid encryption scheme proposed in this paper is shorter and the ciphertext satisfies the verifiability, and the scheme can secure against adaptive chosen ciphertext attacks (CCA2).

This paper presents a comparative assessment of Pole Placement and Linear Quadratic Regulator to control a balancing of two wheels mobile robot. Disturbance is applied to to test the balancing of the robot. The mathematical model of this robot which is highly nonlinear is derived. The final model is then represented in state space form using MATLAB/Simulink application. Simulation on MATLAB application is analyzed and the performance of position, speed, angle, and angle rate of the balancing robot are observed to find out the best controller for two wheels balancing robot model.

Enabling software systems to adjust their protection in continuously changing environments with imperfect context information is a grand challenging problem. The issue of uncertain reasoning based on imperfect information has been overlooked in traditional logic programming with classical negation when applied to dynamic systems. This paper sketches a non-monotonic approach based on Answer Set Programming to reason with imperfect context data in adaptive security where there is little or no knowledge about certainty of the actions and events.

This paper presents comparative study of palmprint matching algorithms for low resolution and noisy images. The principal lines and wrinkles are the only features easily extractable under low resolution. Therefore widely used Radon based approach has been studied and presented in this paper. Based on the computational complexity and fast matching, four different approaches are studied.

Malignant melanomas are the most deadly type of skin cancers however detected early enough give a high chances for successful treatment. The last years saw the dynamic growth of interest of automatic computer-aided skin cancer diagnosis. Every month brings new research results on new approaches to this problem, new methods of preprocessing, new classifiers, new ideas to follow etc. In particular, the rapid development of dermatoscopy, image processing methods, as well as the ever-increasing computing power of computers caused that researchers are able to consider significantly more features of the analyzed lesion than has been done so far using methods recognized in a medical community such as ABCD or Menzies methods. From the other hand more features not always imply an improvement in terms of efficiency of the diagnosis and its transparency. Hence, in this paper we survey the kind of features taken into account by the researchers and then, selected the most efficient set of them. Proposed method jointly selects the optimal set of features representing the analyzed lesion together with the accompanying form of the neural classifier (number of neurons, activation functions). The evolutionary algorithms are used in order to carry out the optimization. Obtained results are even better than the ones obtained by the most efficient these days deep classifiers.

Deep neural networks (DNNs) are successful in many computer vision tasks. However, the most accurate DNNs require millions of parameters and operations, making them energy, computation and memory intensive. This impedes the deployment of large DNNs in low-power devices with limited compute resources. Recent research improves DNN models by reducing the memory requirement, energy consumption, and number of operations without significantly decreasing the accuracy. This paper surveys the progress of low-power deep learning and computer vision, specifically in regards to inference, and discusses the methods for compacting and accelerating DNN models. The techniques can be divided into four major categories: (1) parameter quantization and pruning, (2) compressed convolutional filters and matrix factorization, (3) network architecture search, and (4) knowledge distillation. We analyze the accuracy, advantages, disadvantages, and potential solutions to the problems with the techniques in each category. We also discuss new evaluation metrics as a guideline for future research., Accepted for publication at 2020 IEEE 6th World Forum on Internet of Things (WF-IoT), New Orleans, LA, USA 2020

In recent years, chaos-based image cipher has been widely studied and a growing number of schemes based on permutation-substitution architecture have been proposed. To better meet the challenge of real-time secure image communication applications, this paper suggests a new image encryption scheme using a parallel substitution. In the permutation stage, the Arnold cat map is employed to shuffle the pixel positions so as to erase the strong relationship between adjacent pixels. In the substitution stage, the scrambled image is firstly decomposed into eight bit-planes which are then parallelly mixed with keystreams generated by the chaotic logistic map. Theoretically, the parallel substitution strategy runs eight times faster than the serial strategy on an 8-thread processor as the volume of data processed by each substitution unit is 1/8 of that of the input image. Experimental results show that the proposed parallel scheme runs more than five times faster than the serial scheme. Extensive security analysis is carried out with detailed analysis, demonstrating the satisfactory security of the proposed scheme.

The traditional teaching model of assets evaluation course generally emphasizes the theoretical knowledge impartment, lacking the cultivation of students' ability to solve practical problems. Therefore, it is difficult to satisfy the society needs of people with professional assets evaluation talent. Case teaching is a kind of heuristic teaching method, which places emphasis on the development of the students' spirit of innovation and practical ability. On the basis of characteristics of the assets evaluation course such as theoretical, technical, practical and operational, the implementation of case teaching of assets evaluation course plays a great realistic role in the development of the students' spirit of innovation and practical ability. Taking the characteristics of assets evaluation course for entry points, this paper discusses the meaning and the new teaching model of the case teaching of assets evaluation, providing a probability of wider application of case teaching of assets evaluation.

The Probabilistic Index Map (PIM) model was originally proposed for video processing to extract background of video frames. In this paper, we introduce the PIM model for texture segmentation. We first extract texture features based on Laws and Gabor filters respectively. Then we present a Fuzzy K-Means method to generate the index map and palette, and use the PIM model to improve the segmentation accuracy. Based on the comparison of experimental results produced using different features and different resolutions, we show the proposed method is effective for texture segmentation.

Cerebral aneurysm is a cerebrovascular disorder caused by a weakness in the wall of an artery or vein, which causes a localized dilation or ballooning of the blood vessel. It is life-threatening; hence, an early and accurate diagnosis would be a great aid to medical professionals in making the correct choice of treatment. HemeLB is a massively parallel Lattice-Boltzmann simulation software, which is designed to provide the radiologist with estimates of flow rates, pressures, and shear stresses throughout the relevant vascular structures, intended to eventually permit greater precision in the choice of therapeutic intervention. However, in order to allow surgeries and doctors to view and visualize the results in real-time at medical environments, a cost-efficient, practical platform is needed. In this article, we have developed and evaluated a version of HemeLB on various heterogeneous system-on-chip platforms, allowing users to run HemeLB on a low-cost embedded platform and to visualize the simulation results in real-time. A comprehensive evaluation of implementation on the Zynq SoC and Jetson TX1 embedded graphic processing unit platforms are reported. The achieved results show that the proposed Jetson TX1 implementation outperforms the Zynq implementation by a factor of 19 in terms of site updates per second. 2007-2012 IEEE. Manuscript received February 19, 2019; revised September 20, 2019; accepted November 4, 2019. Date of publication December 2, 2019; date of current version June 3, 2020. This paper was supported by the National Priorities Research Program (NPRP) under Grant 5-792-2-328 from the Qatar National Research Fund (a member of Qatar Foundation). X. Zhai is partially supported by the UK Engineering and Physical Sciences Research Council under Grants EP/R02572X/1 and EP/P017487/1. (Corresponding author: Xiaojun Zhai.) X. Zhai is with the School of Computer Science and Electronic Engineering, University of Essex, CO4 3SQ Colchester, U.K. (e-mail: xzhai@essex.ac.uk). Scopus

Launched in 2018, NASA's Parker Solar Probe (PSP) mission has been described as “The Mission to Touch the Sun”. Over a 7 year prime mission, the spacecraft will make 24 close encounters with the Sun. In order to meet requirements, PSP's telecom subsystem needed to incorporate a Ka-band downlink for radiometrics and science data return. PSP uses a 0.6 meter High Gain Antenna (HGA). This antenna is used only for a Ka-band downlink, and during its use is nominally the only time the science data is transmitted. Aspects of the mission design also restrict the use of the HGA to short periods in between each solar flyby. These constraints require the RF team and Mission Operations team to be extra diligent in optimizing the downlink data return. This paper provides a summary of PSP's first year of operations at Ka-band. In addition to presenting the overall Ka-band performance, other items that will be focused on include: planning considerations for Ka-band coverage with the DSN, observed effectiveness of a HGA calibration, pre-launch link assumptions, and possible future work. With multiple future deep space missions utilizing a Ka-band downlink, this information will hopefully be beneficial to help with planning and optimizing the data return.

Sparse coding is a very powerful method to learn high-level features from raw data input. It is able to learn an over complete basis that has the potential to capture robust and discriminative patterns within the data. However, like many other feature learning algorithms, it is unable to detect very similar features or stimuli on different input channels. In this paper, we propose a novel method to build general features that can be applicable to different sets of channels. This succinct representational model will express the stimuli independent of the locality in which they appeared. As a result, it prepares the groundwork for transferring the learned features from a set of input channels to other possible sets of input channels.

Motor current signature analysis is the reference method for the diagnosis of induction machines. However, it features some drawbacks, as it is bound to steady-state conditions, and fault signatures may be created by different overlapping phenomena. This paper presents a full-time domain-based method for the quantitative evaluation of electrical faults in induction machines. It relies on stator current measurements; thus, it is compliant with switching power converter drives, provided that a proper current measurement technique is available. Moreover, it allows an accurate quantification of faults without requiring spectrum analysis; hence, it achieves high resolution even with a low time acquisition period. A fault index is computed as the energy of the signal in a suitable ldquofault bandwidth,rdquo thus reducing the effects of frequency variations. The aforementioned advantages are perfectly suited to time varying signals and to embedded drive diagnostic systems.

Identifying butterfly species by image patterns is a challenging task in computer vision and pattern recognition community due to many butterfly species having similar shape patterns with complex interior structures and considerable pose variation. In additional, geometrical transformation and illumination variation also make this task more difficult. In this paper, a novel image descriptor, named Gaussian convolution angle (GCA) is proposed for butterfly species classification. The proposed GCA projects the butterfly vein image function and intensity image function along a group of vectors that start from a common contour points and ends at the remaining contour points which results a group of vectors that capture the complex vein patterns and texture patterns of butterfly images. The Gaussian convolutions of different scales are conducted to the resulting vector functions to generate a multiscale GCA descriptors. The proposed GCA is not only invariant to geometrical transformation including rotation, scaling and translation, but also invariant to lighting change. The proposed method has been tested on a publicly available butterfly image dataset that has 832 samples of 10 species. It achieves a classification accuracy of 92.03% which is higher than the benchmark methods.