Search

This paper proposes a novel remote sensing image segmentation method based on Gaussian mixture model and SURF algorithm. Firstly, Gaussian mixture model is used for remote sensing image segmentation. Then the surf matching algorithm is adopted for eliminating misidentified areas. The determinant of Hession matrix (DoH) is used to detect key points in the image. The non-maximum suppression method and interpolation operation are utilized to search and locate the extreme points. The maximum likelihood method is used to estimate model parameters. Some remote sensing images in THE DOTA data set are selected for experimental verification, and the results show that the new algorithm has obvious improvement in segmentation effect and efficiency. In the background complex image segmentation, the improved algorithm has more obvious advantages compared than state-of-the-art segmentation methods.

For the massive large-scale visible image data obtained by satellite, unmanned aerial vehicles, and other reconnaissance platforms, if only relying on manual visual interpretation, there will be problems such as heavy workload, low efficiency, high repeatability, strong subjectivity, and high cost, which cannot meet the demand of modern society for efficient information. Therefore, in order to improve work efficiency, it is necessary to study the rapid automatic region localization in large-scale remote sensing images. That will play an important role in change detection, temperature retrieval, and other files. The development and present situation of the region localization algorithms are analyzed. This paper summarizes the development, improvement, and deficiency of the traditional algorithm, as well as the difficulties and challenges. And the authors make a comparison to the deep learning-based methods. Finally, a possible development direction is prospected.

Aiming at the existing problems of metric-based methods, there are problems such as inadequate feature extraction, inaccurate class feature representation, and single similarity measurement. A new model based on attention mechanism and weight fusion strategy is proposed in this paper. Firstly, the image is passed through the conv4 network with channel attention mechanism and space attention mechanism to obtain the feature map of the image. On this basis, the fusion strategy is used to extract class-level feature representations according to the difference in contributions of different samples to class-level feature representations. Finally, the similarity scores of query set samples are calculated through the network to predict the classification. Experimental results on the miniImageNet dataset and the omniglot dataset demonstrate the effectiveness of the proposed method.

nema

This paper presents a novel web-based crowdsourcing platform for the assessment of the subjective and objective quality of experience (QoE) of the video service in the cloud-server environment. The user has the option to enter subjective QoE data for video service by filling out a web questionnaire. The objective QoE data of the cloud-server, network condition, and the user device is automatically captured by the crowdsourcing platform. Our proposed system collects both objective and subjective QoE simultaneously in real-time. The paper presents the key technologies used in the development of the platform and describes the functional requirements and design ideas of the system in detail. The system collects real-time comprehensive data to enhance the quality of the user experience to provide a valuable reference. The system is tested in a real-time environment and the test results are given in terms of the system performance. The crowdsourcing platform has new features of real-time network monitoring, the client device, and cloud monitoring, which currently has not been provided by existing web platforms and crowdsourcing frameworks. The results show that 1MB buffer is filled 100% very soon after starting watching videos from the crowdsourcing platform.

According to the research works of the references on chaotic cryptanalysis, many recent chaotic image encryption algorithms cannot resist chosen-plaintext attacks. Although some chaotic image encryption algorithms introduce plain-image information, they still violate some design requirements of modern cryptosystems. The traditional image encryption algorithm has the disadvantages of low security and time-consuming. Therefore, after analyzing many literatures related to chaotic image encryption and aiming to improve the security of ciphertext and effectively resist plaintext attack, we propose a hybrid chaotic system-oriented artificial fish swarm neural network method for image encryption. Firstly, based on the nonlinear combination theory, the Logistic, Tent and Sine mappings are used to design the hybrid complex chaotic system. According to the pixel value of plaintext, it generates the initial value to output the chaotic sequence. Taking a chaotic sequence as the input layer, the artificial neural network is introduced to train and learn the chaotic sequence. It eliminates the chaotic periodicity and outputs the neural network sequence. The artificial fish swarm algorithm is used to optimize the initial weight and threshold of the Elman neural network. The set obfuscation method is defined to scramble the plaintext. A quantization method is constructed to quantify the neural network sequence and obtain the key flow. The mixed images are classified, the key flow is combined, the pixel value is changed and the ciphertext is output by designing the piecewise diffusion technology. Experimental results show that compared with the current state-of-the-art image encryption methods, the proposed algorithm has higher security and resist plaintext attack ability.

Anatomical maps depicted by medical images reflect the health of the human body. Understanding the detailed division of the regions in the anatomical maps is helpful for diagnosis and further treatment. For example, the lung window contains the lung parenchyma, mediastinum, and diseased tumor. Accurate tumor segmentation can be more comprehensive and accurate to determine the disease condition to select the appropriate radiotherapy means to improve the success rate of treatment. However, the traditional manual segmentation task is large, and the segmentation results vary with the experience level of the observer. Therefore, the development of automatic and reliable segmentation methods is a feasible and important task in clinical practice. Therefore, this paper proposes a multichannel convolutional neural network (CNN)-based fuzzy active contour model for medical image segmentation. Medical images are gridded and labeled using the superpixel SLIC method. The gridded images are used as data sets to train the multichannel CNN network to segment superpixels of organ boundaries, and the seed points of these superpixels are connected to form rough segmentation boundaries. The rough segmentation boundary is used as the initial contour, and the organ boundary in the medical image is obtained by fuzzy active contour segmentation. Finally, we conduct comparison experiments with the state-of-the-art segmentation methods, the results of the subjective evaluation and objective evaluation show that the proposed method has a better segmentation effect and robustness.

Aiming at facial expression recognition, the recognition rate is not high due to noise and occlusion. A hybrid approach of facial expression has been presented by combining local and global features. First, feature extraction is performed to fuse the histogram of oriented gradients (HOG) descriptor with the compounded local ternary pattern (C-LTP) descriptor. Second, features extracted by HOG and C-LTP are fused into a single feature vector. Third, the feature vector is sent to a multi-class support vector machine classifier for facial classification. Finally, the proposed method is compared with the existing facial expression recognition methods in three public facial expression image databases, and the results show that the recognition rates of the proposed method in MMI, JAFFE and CK[KG-*2]+ databases are 98.28%, 95.75% and 99.64%, respectively. The average recognition rate is 10% higher than other methods, which is better than other existing methods. The results of this study provide a reference for the research of facial expression recognition in many situations. The method of facial expression recognition proposed can effectively promote the development of human-computer interaction system and the study of computer image understanding. It is of great significance to realize the fusion of human language and natural language, as well as the establishment and implementation of the connection model between language and expression.

Abstract: Digital well-being records are multimodal and high-dimensional (HD). Better theradiagnostics stem from new computationally thorough and edgy technologies, i.e., hyperspectral (HSI) imaging, super-resolution, and nanoimaging, but advance mess data portrayal and retrieval. A patient's state involves multiple signals, medical imaging (MI) modalities, clinical variables, dialogs between clinicians and patients, metadata, genome sequencing, and signals from wearables. Patients' high volume, personalized data amassed over time have advanced artificial intelligence (AI) models for higherprecision inferences, prognosis, and tracking. AI promises are undeniable, but with slow spreading and adoption, given partly unstable AI model performance after real-world use. The HD data is a ratelimiting factor for AI algorithms generalizing real-world scenarios. This paper studies many health data challenges to robust AI models' growth, aka the dimensionality curse (DC). This paper overviews DC in the MIs' context, tackles the negative out-of-sample influence and stresses important worries for algorithm designers. It is tricky to choose an AI platform and analyze hardships. Automating complex tasks requires more examination. Not all MI problems need automation via DL. AI developers spend most time refining algorithms, and quality data are crucial. Noisy and incomplete data limits AI, requiring time to handle control, integration, and analyses. AI demands data mixing skills absent in regular systems, requiring hardware/software speed and flexible storage. A partner or service can fulfill anomaly detection, predictive analysis, and ensemble modeling.

Traditional object detection algorithms and strategies are difficult to meet the requirements of data processing efficiency, performance, speed and intelligence in object detection. Through the study and imitation of the cognitive ability of the brain, deep learning can analyze and process the data features. It has a strong ability of visualization and becomes the mainstream algorithm of current object detection applications. Firstly, we have discussed the developments of traditional object detection methods. Secondly, the frameworks of object detection (e.g. Region-based CNN (R-CNN), Spatial Pyramid Pooling Network (SPP-NET), Fast-RCNN and Faster-RCNN) which combine region proposals and convolutional neural networks (CNNs) are briefly characterized for optical remote sensing applications. You only look once (YOLO) algorithm is the representative of the object detection frameworks (e.g. YOLO and Single Shot MultiBox Detector (SSD)) which transforms the object detection into a regression problem. The limitations of remote sensing images and object detectors have been highlighted and discussed. The feasibility and limitations of these approaches will lead researchers to prudently select appropriate image enhancements. Finally, the problems of object detection algorithms in deep learning are summarized and the future recommendations are also conferred.

Network intrusion detection (NID) is an important method for network system administrators to detect various security holes. The performance of traditional NID methods can be affected when unknown or new attacks are detected. Compared with other machine learning methods, the intrusion detection method based on convolutional neural network (CNN) can significantly improve the accuracy of classification, but the convergence speed and generalization ability of CNN are not ideal in model training process resulting in a low true rate and a high false alarm rate. To solve the above problems, this paper proposes a deep multi-scale convolutional neural network (DMCNN) for network intrusion detection. Different levels of features in a large number of high-dimensional unlabeled original data are extracted by different scales convolution kernel. And the learning rate of network structure is optimized by batch normalization method to obtain the optimal feature representation of the raw data. We use NSL-KDD dataset as the benchmark thus we can compare the performance of our proposed method with other existing works. This dataset includes two testing sets which are the first one is KDDTest+ while the second one is $$\text {KDDTest}^{-21}$$ which is more difficult to be classified. The experimental results reveal that the AC and TPR are higher through our DMCNN model. Especially, in terms of DOS, the AC appropriately reaches to 98%. DMCNN has a high intrusion detection accuracy and a low false alarm rate, which overcomes the limitations of using the traditional detection methods and makes the new approach an attractive one for practical intrusion detection.

In the transmission of medical images, if the image is not processed, it is very likely to leak data and personal privacy, resulting in unpredictable consequences. Traditional encryption algorithms have limited ability to deal with complex data. The chaotic system is characterized by randomness and ergodicity, which has advantages over traditional encryption algorithms in image encryption processing. A novel V-net convolutional neural network (CNN) based on four-dimensional hyperchaotic system for medical image encryption is presented in this study. Firstly, the plaintext medical images are processed into 4D hyperchaotic sequence images, including image segmentation, chaotic system processing, and pseudorandom sequence generation. Then, V-net CNN is used to train chaotic sequences to eliminate the periodicity of chaotic sequences. Finally, the chaotic sequence image is diffused to change the raw image pixel to realize the encryption processing. Simulation test analysis demonstrates that the proposed algorithm has better effect, robustness, and plaintext sensitivity.

Currently, medical image segmentation has attracted more attention from researchers, which can assist in medical diagnosis. However, in the process of traditional medical image segmentation, it is sensitive to the initial contour and noise, which is difficult to deal with the weak edge image, complex iterative process. In this paper, we propose a new medical image segmentation method, which adopts density-oriented BIRCH (balanced iterative reducing and clustering using hierarchies) clustering method to modify active contour model and improve the robustness of noise. The BIRCH is a multi-stage clustering method using clustering feature tree. The improved model can effectively deal with the gray non-uniformity of real medical images. And we also introduce a new energy function in active contour model to make the contour curve approach to the edge, and finally stay at the edge of the image to complete the object segmentation. Experimental results show that this new model can overcome the influence of complex background on medical image segmentation and improve the speed and accuracy of medical segmentation results.

Airport detection in remote sensing images is an important process which plays a significant role in military and civil areas. Mostly, conventional algorithms have been used for airport detection f...

To address the low accuracy problem of playground detection under complex background, the accurate playground localization based on multi-feature extraction and cascade classifier is proposed in th...

Hot region selection (HRS) is usually performed for larger scenes, especially in military target detection and battlefield situational awareness applications. Many state-of-the-art methods pay more attention on object detection. Although they can detect the region of interest in one image, these images are highly affected by illumination, rotation, and scale changes, which further increases the complexity of analysis compared to those obtained using standard remote sensing platforms. The study of HRS is of great importance in the analysis of remote sensing images. The current research focuses on the specific type of object area detection. A well-developed HRS needs to have three properties: uniform highlighting of the entire HRS, well-defined boundaries, and good robustness. Motivated by these requirements, a HRS method based on the selective search method and modified fuzzy c-means (FCM) in remote sensing images is proposed to address the detection of potential hot regions in large-scale remote sensing images. First, we create a Gaussian curvature filter to preprocess large scale remote sensing images. Second, a modified FCM segmentation method is utilized to segment the image. Third, an enhanced selective search method is adopted to establish well-defined boundaries for the HRS and to improve the immunity to noise. The geographic information is presented in this phase, which is conducive to improve the detection accuracy. In the experimental section, we compare our new method with four other extraction models on three data sets. The experimental results show that compared to the other competing models, the new model better defines the hot regions and obtains more entire boundaries in terms of Overlap and mA P.

With the rapid development of social network, intelligent terminal and automatic positioning technology, location-based social network (LBSN) service has become an important and valuable application. Point of interest (POI) recommendation is an important content in LBSN, which aims to recommend new locations of interest for users. It can not only alleviate the information overload problem faced by users in the era of big data, improve user experience, but also help merchants quickly find target users and achieve accurate marketing. Most of the works are based on users' check-in history and social network data to model users' personalized preferences for interest points, and recommend interest points through collaborative filtering and other recommendation technologies. However, in the check-in history, the multi-source heterogeneous information (including the position, category, popularity, social, reviews) describes user activity from different aspects which hides people's life style and personal preference. However, the above methods do not fully consider these factors' combined action. Considering the data privacy, it is difficult for individuals to share data with others with similar preferences. In this paper, we propose a privacy protection point of interest recommendation algorithm based on multi-exploring locality sensitive hashing (LSH). This algorithm studies the POI recommendation problem under distributed system. This paper introduces a multi-exploring method to improve the LSH algorithm. On the one hand, it reduces the number of hash tables to decrease the memory overhead; On the other hand, the retrieval range on each hash table is increased to reduce the time retrieval overhead. Meanwhile, the retrieval quality is similar to the original algorithm. The proposed method uses modified LSH and homomorphic encryption technology to assist POI recommendation which can ensure the accuracy, privacy and efficiency of the recommendation algorithm, and it verifies feasibility through experiments on real data sets. In terms of root mean square error (RMSE), mean absolute error (MAE) and running time, the proposed method has a competitive advantage.

Image segmentation (Lin et al. 2018, Shoulin et al. 2018) is the basic link of image processing and machine vision, so it is very important to study this content. Because the image can be interfere...

Traditional image segment algorithms have some demerits: edge blur, discontinuous edge is susceptible to peripheral irrelevant information, Poor noise resistance. Therefore, we propose an improved ...

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.

With the rapid development of social network, intelligent terminal and automatic positioning technology, location-based social network (LBSN) service has become an important and valuable application. Point of interest (POI) recommendation is an important content in LBSN, which aims to recommend new locations of interest for users. It can not only alleviate the information overload problem faced by users in the era of big data, improve user experience, but also help merchants quickly find target users and achieve accurate marketing. Most of the works are based on users' check-in history and social network data to model users' personalized preferences for interest points, and recommend interest points through collaborative filtering and other recommendation technologies. However, in the check-in history, the multi-source heterogeneous information (including the position, category, popularity, social, reviews) describes user activity from different aspects which hides people's life style and personal preference. However, the above methods do not fully consider these factors' combined action. Considering the data privacy, it is difficult for individuals to share data with others with similar preferences. In this paper, we propose a privacy protection point of interest recommendation algorithm based on multi-exploring locality sensitive hashing (LSH). This algorithm studies the POI recommendation problem under distributed system. This paper introduces a multi-exploring method to improve the LSH algorithm. On the one hand, it reduces the number of hash tables to decrease the memory overhead; On the other hand, the retrieval range on each hash table is increased to reduce the time retrieval overhead. Meanwhile, the retrieval quality is similar to the original algorithm. The proposed method uses modified LSH and homomorphic encryption technology to assist POI recommendation which can ensure the accuracy, privacy and efficiency of the recommendation algorithm, and it verifies feasibility through experiments on real data sets. In terms of root mean square error (RMSE), mean absolute error (MAE) and running time, the proposed method has a competitive advantage.

Object detection in optical remote sensing imagery is being explored to deal with arbitrary orientations and complex appearance which is still a major issue in recent years. To perceive a better solution to the addressed problem, the post-processing of bounding boxes (BBs) has been evaluated and discussed for the applications of object detection. In this paper, the proposed method has divided into two stages; the first stage is based on thresholding of BBs with respect to the confidence values and the second stage is based on the area-based BB regression (BBR). In BBR, the area of each BB was estimated then the oversized and undersized BBs were removed with respect to the size of objects which are being detected. The widely known region-based approaches RCNN, Fast-RCNN and Faster-RCNN are used for evaluation and comparative analysis validates the proposed framework. The results show that the proposed post-processing is very effective for each kind of region-based detector.

The extensive acquisition and using of high-resolution remote sensing images have greatly promoted the development of airport detection. However, due to the complex shape, background and different scale of the airport location, the real-time and accuracy of airport detection are also facing major challenges. These factors can reduce the detection accuracy. To solve the above problems, we propose an improved faster region-based convolutional neural network (RCNN) detection method for airport detection in large scale remote sensing images. Multi-scale training is applied to faster RCNN to enhance the robustness of network for detecting airport with different sizes. Meanwhile, we adopt the modified multitask loss function to improve the accuracy of airport detection. Online hard example mining strategy is introduced to decrease the redundant negative samples in the training process. Then the non-maximum suppression method is used to remove the redundant boxes of the detected airport. Finally, we conduct sufficient experiments with the airport data obtained from Google Earth and make comparison with the state-of-the-art airport detection methods. The results show that the proposed method can accurately detect different airports under complex background with high detection rate, low false alarm rate and short running time.

Image fusion is an important part of image processing because it can collaboratively use image information from different sensors for the same scene. Image fusion outputs the images, which are more...

Network intrusion can enter the network through informal channels. Some illegal users utilize Trojans and self-programmed attack to change the network security system, so that the system loses the defense and alarm function and the Hacker can steal the internal information. Network intrusion seriously harms the security of network information and the legitimate rights of users. Therefore, a bagging strategy-based kernel extreme learning machine for complex network intrusion detection is presented in this paper. This method adopts a bagging strategy to train several sub-kernel extreme learning machines independently. Then the integrated gain of above machines is measured based on the margin distance minimization (MDM) criterion. Selected machines with high gain degree are selected for selective integration to obtain selective integrated learners with strong generalization ability and high efficiency. Then an improved universal gravitation search algorithm is used to optimize the kernel parameters. Meanwhile, a sub-kernel extreme learning machine online update strategy based on incremental learning of batch samples is introduced to realize the online update of intrusion detection model, so that the proposed detection method can effectively be adapted to the changes of complex network environment. Finally, experiments illustrate that the proposed method has better effect on network intrusion detection in terms of detection accuracy and speed, especially for unknown network intrusion connection events, the response speed is fast, the false alarm rate is low.

As we all know, The feature obtained by the convolutional neural network (CNN) is the expression of higher level for the input image, which can improve the discrimination of the input data. In this paper, we use an excellent convolutional neural network framework to design the pedestrian detection algorithm. Aiming at the background interference between the descendant and the occlusion between the pedestrians, the head of pedestrian is considered as the detection object, so the features of the object to be tested are reduced. To obtain a better detection effect and accelerate the calculation speed, the regional proposal network (RPN) in fast RCNN can be used as a better pedestrian detector, but the latter classifier reduces the performance. The object detection network is improved based on fast RCNN, and the features of the first layer are merged, so that the obtained features are richer. Experiments show that the new method can separate the objects from the background accurately.

Recent research progress in shadow detection has leveraged the development of remote sensing and computer vision. Since shadows of buildings, trees, bridges in one image can provide useful information about the scene to help people understand the shape, feature or estimate their locations and orientations of original objects, especially for damaged objects. In this study, a novel shadow detection algorithm for remote sensing imagery, called self-supervised learning method is proposed. The aim of this work is to generate shadow ratio threshold automatically without human interaction. To alleviate the traditional drawbacks of shadow detection, we fully combine supervised and unsupervised shadow detection method to suggest a self-supervised learning method, which supports us a strongly clue with establishing the relation of shadow and its original object. Subsequently, we benefit from gray-scale histogram to extract shadow segments, then shadow outlines are obtained. Finally, we assess the shadow detection performance of the proposed approach by comparing our results with the state-of-the-art methods. The results reveal the applicability and precision of the proposed self-supervised learning shadow detection technique.

The original BP neural network has some disadvantages, such as slow convergence speed, low precision, which is easy to fall into local minimum value. So this paper proposes an improved particle swarm optimization (PSO) algorithm to optimize BP neural network. In this new algorithm, PSO uses improved adaptive acceleration factor and improved adaptive inertia weight to improve the initial weight value and threshold value of BP neural network. And we give the detailed improved process. At the end, simulation results show that the new algorithm can improve convergence rate and precision of prediction of BP neural network, which reduces the error of prediction. At the end, we use multimedia evaluation model to verify the new method’s performance.

Region proposals are very important for several perfections of remote sensing applications such as vehicle detection, traffic surveillance and intelligent transport system. In this paper, an efficient region proposal approach has been proposed. The framework is organized into two key steps. The first step is based on extracting region proposals using Cascade system. The second step is based on the classification of extracted region proposals which is performed by transfer learning using Convolutional Neural Networks (CNN) and AlexNet architecture is utilized for transfer learning. The aim of this investigation is to evaluate the proposed method for vehicle detection. The selective search (SS) method is also briefly discussed for comparison. The results regarding vehicle detection are very promising.

In order to solve the problem that the semantic gap cannot be alleviated effectively in the field of medical image retrieval, we propose a new graph semi-supervised learning method for medical image automatic annotation. Because the model adopts semi-supervised technology, it can learn from abundant unlabeled instances to avoid the decreasing of generalization ability which is induced by the relative lack of labeled data. Furthermore, by improving graph based semi-supervised learning technology with normalization and modification of decision boundary on its iterative results, the scoring model effectively reduces the bad impact of asymmetric dataset. In view of the relationship between word extraction and image in image learning model, we analyze image similarity calculation in detail. It effectively combines together into the physician's diagnosis information as high-level semantic feature of image, to calculate the similarity between images more effectively. Finally, the Toy data and clinical data sets gastroscope image sets are conducted with a series of experiments, the results show that the method is superior to traditional image annotation method in this paper.

The using of deep learning method belongs to the application and research of artificial intelligence technology in medical field for assisting medical image information processing. Due to the larger difference between the image underlying character features and high-level semantic concepts described, the existing image annotation algorithms perform not very ideal, therefore, the we propose an image annotation framework based on multimodule deep transfer learning method, aiming at the application of deep convolution neural network technology to optimize the neural network parameters, and improve the tagging accuracy in detail. The proposed multimodule deep learning annotation framework uses two stages to optimiz neural network parameters: (1) using deep neural network to optimize the single-mode condition parameters; (2) using the phase correlation, to realize the multi-mode state, the experiment of the optimal combination of public data sets shows that the scheme can effectively improve the performance of image annotation.

Data stream sliding window is a common query method, but traditional approaches can get inaccurate data results. Due to the blocked character of join and the infinite character of data stream, there must be some constraints on join operations. In order to solve these constraints, this paper proposes a strategy based on load shedding techniques for sliding window aggregation queries over data stream for basic query processing and optimization. The new strategy supports multiple streams and multiple queries. Through experimental tests, the results show that the new strategy based on load shedding techniques can greatly improve query processing efficiency. Finally, we make a comparison to other join query strategies over data stream to verify the new method. Join query strategy over data stream based on sliding windows is an effective method, which can effectively reduce query processing time.

Currently, big data is a new and hot issue. Particularly, the rapid growth of the Internet of Things causes a sharp growth of data. Enormous amounts of networking sensors are continuously collecting and transmitting data to be stored and processed in the cloud, including remote sensing data, environmental data, and geographical data. And region is regarded as the very important object in remote sensing data, which is mainly researched in this article. Region search is a crucial task in remote sensing process, especially for military area and civilian fields. It is difficult to fast search region accurately and achieve generalizability of the regions’ features due to the complex background information, as well as the smaller size. Especially, when processing region search in large-scale remote sensing image, detailed information as the feature can be extracted in inner region. To overcome the above difficulty region search task, we propose an accurate and fast region search in optical remote sensing images under cloud computing environment, which is based on hybrid convolutional neural network. The proposed region search method partitioned into four processes. First, fully convolutional network is adopted to produce all the candidate regions that contain the possible object regions. This process avoids exhaustive search for input images. Then, the features of all candidate regions are extracted by a fast region-based convolutional neural network structure. Third, we design a new difficult sample mining method for the training process. At the end, in order to improve the region search precision, we use an iterative bounding box regression algorithm to normalize the detected bounding boxes, in which the regions contain candidate objects. The proposed algorithm is evaluated on optical remote sensing images acquired from Google Earth. Finally, we conduct the experiments, and the obtained results show that the proposed region search method constantly achieves better results regardless of the type of images tested. Compared with traditional region search methods, such as region-based convolutional neural network and newest feature extraction frameworks, our proposed methods show better robustness with complex context semantic information and backgrounds.

Though many intelligence algorithms are used for travelling salesman problem (TSP), the main objective of this paper is to execute new approach to obtain significant improvements. This paper proposes an improved wolf colony search algorithm based on search strategy. First, we introduce interaction strategy into travel behaviour and calling behaviour to promote the communication between artificial wolves, which can improve the information acquirement for wolves and enhance the exploring ability of wolves. Second, we present adaptive siege strategy for siege behaviour, which guarantees that the new algorithm can obtain better collaborative search feature. Therefore, the range of wolf siege constantly decreases and the mining ability of wolf algorithm increases with the new strategy. Finally, experiments are carried out to verify the effectiveness of new method compared with other algorithms for TSP problems. The results show that the improved wolf colony search algorithm has higher solving accuracy, faster convergence speed.

User interface (UI) is an interaction and information exchange medium between systems and users, which is convenient to communicate and enable users to operate hardware easily and effectively to achieve the bidirectional interaction. Traditional UI has difficulty in satisfying requirements of users. Therefore, this paper proposes a new intelligence interface scheme based on K-means gaze tracking. First, it uniformly describes the user, interface and system in intelligent interface interaction framework based on visual attention selection mechanism. Second, it uses K-means method to calculate attention degree value of interface on the basis of mapping relation of user, interface and system. Third, it adopts visual attention allocation strategy to predict users' interest degree on interface. Finally, experiments are conducted to verify the performance of our new scheme. The results show that the accuracy of intelligent interface predicting user decision intention is very high. This new method is a kind of selectable solution for marking intelligent interface of user interest goal automatically. What's more, intelligence interface scheme based on K-means gaze tracking can effectively improve the quality of gaze tracking.

Antenna miniaturization plays an important role in modern ultrashort wave communication system. Monopole antenna has the advantages with simple construction, feed easily, small lateral size and omnidirectional radiation. Monopole antenna miniaturization means that it has smaller physical size, smaller wind resistance and lower cost. Antenna has a better concealment and strong durability. However, it ensures the bandwidth and gain of the antenna at the same time which is very difficult. So studying the monopole antenna miniaturization is necessary. The antenna miniaturization is beneficial to improve its concealment and stability, hence, it is very important to design the antenna miniaturization. In this paper, we propose time-domain integration method based on moments and apply it into ultra wideband response of monopole antenna. This numerical algorithm adopts Gail polynomial as expanding basis function in the whole time domain space to expand time domain coefficients of unknown current. Also, it makes numerical simulation for the transient state response of monopole antenna and makes comparison with testing data in experiments. The new method matches experimental data well. Finally, we make a analysis for our scheme to verify its effectiveness. The analysis method can provide input conditions for the response of electromagnetic interference by antenna or system on purpose. What's more, we will study more new methods to design better monopole antenna in the future.

Traditional distinction and recognition methods of electromagnetic radiation source have some shortcomings with highly error and long time. Therefore, Wilson algorithm, Lee algorithm and near field wave impedance theory have been proposed to analyze radiation source characteristics. And it presented common-mode radiation noise testing method of poor-average algorithm and differential mode noise testing method of variance algorithm respectively according to the different radiation source characteristics. Nevertheless, there are still some questions with imprecise location. By acquiring electric field intensity distribution around space of electromagnetic radiation sources, we can determine the direction characteristics of the radiation sources. Using the directional characteristic parameter with high differentiation degree, we make accurate classification and recognition for electromagnetic radiation sources based on support vector machine (SVM). In this paper, we analyze three basic antenna radiation models. It establishes cube reception array to obtain its filed strength value of far zone field in the same position, in addition, it adopts support vector machine method to deal with data and sets up a evaluation model to analyze recognition accuracy. Finally, experimental results show that the proposed method has high recognition accuracy. What's more, we illustrate the new scheme from three aspects with anti-noise performance, data normalization method and F1 value.

TEM horn antenna is a kind of ultra-wide band antennas. Ultra-wide band antenna can make a pulse signal have minimum distortion through radiation, namely the smaller the impulse response trailing is. TEM horn antenna radiate TEM spherical wave. It is limited by center of antenna port and the edge phase difference, so generally it requires a small horn flare angle. There are many loads in the TEM horn antenna to reduce low frequency reflection and absorb some low frequency energy. TEM horn antenna with a small field angle can be approximated as electrical dipole antenna in a low frequency. And loading low frequency compensation circuits can constitute a magnetic dipole, so electric dipole and magnetic dipole forms a combination oscillator antenna, which can improve the low frequency radiation performance of antenna. Taking power capacity and caliber match into consideration, we design a new TEM horn antenna based on ultra-wide band broken line. Compared to traditional constant impedance TEM horn antenna, the new design has better performance, on account for the fact that the new TEM horn antenna adopts transitional feed-in structure in coaxial panel, and the radiation plate is separated into two broken line forms. By theoretical simulation and experimental research, we compare and analyze the input, radiation characteristics of two kinds of antenna. Results indicate that when width of Gaussian pulse excitation is 960 ps, the transition peak power transmission efficiency of coaxial panel will reach at 85%. Peak electric field radiation of new TEM horn antenna improves by 30% than constant impedance horn. Main lobe width of H aside pattern decreases to 60° noticeably compared with original 80°. In addition, it perfects transmission radiation characteristics in 700MHz band, transient pulse peak power capacity comes to 15GW.