To alleviate the expensive human labeling problem, semi-supervised semantic segmentation utilizes a few labeled images along with an abundance of unlabeled images to predict the pixel-level label maps with the same size. Previous methods often rely on co-training with two convolutional networks with the same architecture but different initialization, which fails to capture sufficiently diverse features. This limitation motivates us to employ tri-training and design a triple-view encoder to utilize encoders with different architectures to derive diverse features, while leveraging knowledge distillation to capture complementary semantics among these encoders. Moreover, existing approaches simply concatenate features from both encoder and decoder, and the simple concatenation requires a large memory cost. This inspires us to present a dual-frequency decoder that selects those important features by projecting the spatial-domain features into the frequency domain, where a dual-frequency channel attention mechanism is applied to evaluate the feature importance. Therefore, we propose a Triple-view Knowledge Distillation framework, termed TriKD, for semi-supervised semantic segmentation. It comprises the triple-view encoder and the dual-frequency decoder. Extensive experiments conducted on two benchmarks,i.e., Pascal VOC 2012 and Cityscapes, validate the superiority of our method, achieving a satisfying tradeoff between precision and inference speed. Our code is available at GitHub.

Video object segmentation has been applied to various computer vision tasks, such as video editing, autonomous driving, and human-robot interaction. However, the methods based on deep neural networks are vulnerable to adversarial examples, which are the inputs attacked by almost human-imperceptible perturbations, and the adversary (i.e., attacker) will fool the segmentation model to make incorrect pixel-level predictions. This will rise the security issues in highly-demanding tasks because small perturbations to the input video will result in potential attack risks. Though adversarial examples have been extensively used for classification, it is rarely studied in video object segmentation. Existing related methods in computer vision either require prior knowledge of categories or cannot be directly applied due to the special design for certain tasks, failing to consider the pixel-wise region attack. Hence, this work develops an object-agnostic adversary that has adversarial impacts on VOS by first-frame attacking via hard region discovery. Particularly, the gradients from the segmentation model are exploited to discover the easily confused region, in which it is difficult to identify the pixel-wise objects from the background in a frame. This provides a hardness map that helps to generate perturbations with a stronger adversarial power for attacking the first frame. Empirical studies on three benchmarks indicate that our attacker significantly degrades the performance of several state-of-the-art video object segmentation models.

Video prediction is a pixel-level task that generates future frames by employing the historical frames. There often exist continuous complex motions, such as object overlapping and scene occlusion in video, which poses great challenges to this task. Previous works either fail to well capture the long-term temporal dynamics or do not handle the occlusion masks. To address these issues, we develop the fully convolutional Fast Fourier Inception Networks for video prediction, termedFFINet, which includes two primary components, i.e., the occlusion inpainter and the spatiotemporal translator. The former adopts the fast Fourier convolutions to enlarge the receptive field, such that the missing areas (occlusion) with complex geometric structures are filled by the inpainter. The latter employs the stacked Fourier transform inception module to learn the temporal evolution by group convolutions and the spatial movement by channel-wise Fourier convolutions, which captures both the local and the global spatiotemporal features. This encourages generating more realistic and high-quality future frames. To optimize the model, the recovery loss is imposed to the objective, i.e., minimizing the mean square error between the ground-truth frame and the recovery frame. Both quantitative and qualitative experimental results on five benchmarks, including Moving MNIST, TaxiBJ, Human3.6 M, Caltech Pedestrian, and KTH, have demonstrated the superiority of the proposed approach.

Video summarization has established itself as a fundamental technique for generating compact and concise video, which alleviates managing and browsing large-scale video data. Existing methods fail to fully consider the local and global relations among frames of video, leading to a deteriorated summarization performance. To address the above problem, we propose a graph convolutional attention network (GCAN) for video summarization. GCAN consists of two parts, embedding learning and context fusion, where embedding learning includes the temporal branch and graph branch. In particular, GCAN uses dilated temporal convolution to model local cues and temporal self-attention to exploit global cues for video frames. It learns graph embedding via a multi-layer graph convolutional network to reveal the intrinsic structure of frame samples. The context fusion part combines the output streams from the temporal branch and graph branch to create the context-aware representation of frames, on which the importance scores are evaluated for selecting representative frames to generate video summary. Experiments are carried out on two benchmark databases, SumMe and TVSum, showing that the proposed GCAN approach enjoys superior performance compared to several state-of-the-art alternatives in three evaluation settings.

Summary The abnormal visual event detection is an important subject in Smart City surveillance where a lot of data can be processed locally in edge computing environment. Real‐time and detection effectiveness are critical in such an edge environment. In this paper, we propose an abnormal event detection approach based on multi‐instance learning and autoregressive integrated moving average model for video surveillance of crowded scenes in urban public places, focusing on real‐time and detection effectiveness. We propose an unsupervised method for abnormal event detection by combining multi‐instance visual feature selection and the autoregressive integrated moving average model. In the proposed method, each video clip is modeled as a visual feature bag containing several subvideo clips, each of which is regarded as an instance. The time‐transform characteristics of the optical flow characteristics within each subvideo clip are considered as a visual feature instance, and time‐series modeling is carried out for multiple visual feature instances related to all subvideo clips in a surveillance video clip. The abnormal events in each surveillance video clip are detected using the multi‐instance fusion method. This approach is verified on publically available urban surveillance video datasets and compared with state‐of‐the‐art alternatives. Experimental results demonstrate that the proposed method has better abnormal event detection performance for crowded scene of urban public places with an edge environment.

Big sensing data is commonly encountered from various surveillance or sensing systems. Sampling and transferring errors are commonly encountered during each stage of sensing data processing. How to recover from these errors with accuracy and efficiency is quite challenging because of high sensing data volume and unrepeatable wireless communication environment. While Cloud provides a promising platform for processing big sensing data, however scalable and accurate error recovery solutions are still need. In this paper, we propose a novel approach to achieve fast error recovery in a scalable manner on cloud. This approach is based on the prediction of a recovery replacement data by making multiple data sources based approximation. The approximation process will use coverage information carried by data units to limit the algorithm in a small cluster of sensing data instead of a whole data spectrum. Specifically, in each sensing data cluster, a Euclidean distance based approximation is proposed to calculate a time series prediction. With the calculated time series, a detected error can be recovered with a predicted data value. Through the experiment with real world meteorological data sets on cloud, we demonstrate that the proposed error recovery approach can achieve high accuracy in data approximation to replace the original data error. At the same time, with MapReduce based implementation for scalability, the experimental results also show significant efficiency on time saving.

Recommendation technology is an important part of the Internet of Things (IoT) services, which can provide better service for users and help users get information anytime, anywhere. However, the traditional recommendation algorithms cannot meet user's fast and accurate recommended requirements in the IoT environment. In the face of a large-volume data, the method of finding neighborhood by comparing whole user information will result in a low recommendation efficiency. In addition, the traditional recommendation system ignores the inherent connection between user's preference and time. In reality, the interest of the user varies over time. Recommendation system should provide users accurate and fast with the change of time. To address this, we propose a novel recommendation model based on time correlation coefficient and an improved K-means with cuckoo search (CSK-means), called TCCF. The clustering method can cluster similar users together for further quick and accurate recommendation. Moreover, an effective and personalized recommendation model based on preference pattern (PTCCF) is designed to improve the quality of TCCF. It can provide a higher quality recommendation by analyzing the user's behaviors. The extensive experiments are conducted on two real datasets of MovieLens and Douban, and the precision of our model have improved about 5.2 percent compared with the MCoC model. Systematic experimental results have demonstrated our models TCCF and PTCCF are effective for IoT scenarios.

Tensor data (i.e., the data having multiple dimensions) are quickly growing in scale in many practical applications, which poses new challenges for data modeling and analysis approaches, such as high-order relations of large complexity, gross noise, and varying data scale. Existing low-rank data analysis methods, which are effective at analyzing matrix data, may fail in the regime of tensor data due to these challenges. A robust and scalable low-rank tensor modeling method is heavily desired. In this paper, we develop an online robust low-rank tensor modeling (ORLTM) method to address these challenges. The ORLTM method leverages the high-order correlations among all tensor modes to model an intrinsic low-rank structure of streaming tensor data online and can effectively analyze data residing in a mixture of multiple subspaces by virtue of dictionary learning. ORLTM consumes a very limited memory space that remains constant regardless of the increase of tensor data size, which facilitates processing tensor data at a large scale. More concretely, it models each mode unfolding of streaming tensor data using the bilinear formulation of tensor nuclear norms. With this reformulation, ORLTM employs a stochastic optimization algorithm to learn the tensor low-rank structure alternatively for online updating. To capture the final tensors, ORLTM uses an average pooling operation on folded tensors in all modes. We also provide the analysis regarding computational complexity, memory cost, and convergence. Moreover, we extend ORLTM to the image alignment scenario by incorporating the geometrical transformations and linearizing the constraints. Extensive empirical studies on synthetic database and three practical vision tasks, including video background subtraction, image alignment, and visual tracking, have demonstrated the superiority of the proposed method.

The widespread use of smartphones has brought great convenience to our daily lives, while at the same time we have been increasingly exposed to security threats. Keystroke security is an essential element in user privacy protection. In this paper, we present GazeRevealer, a novel side-channel based keystroke inference framework to infer sensitive inputs on smartphone from video recordings of victim's eye patterns captured from smartphone front camera. We observe that eye movements typically follow the keystrokes typing on the number-only soft keyboard during password input. By exploiting eye patterns, we are able to infer the passwords being entered. We propose a novel algorithm to extract sensitive eye pattern images from video streams, and classify different eye patterns with Support Vector Classification. We also propose a novel enhanced method to boost the inference accuracy. Compared with prior keystroke detection approaches, GazeRevealer does not require any external auxiliary devices, and it relies only on smartphone front camera. We evaluate the performance of GazeRevealer with three different types of smartphones, and the result shows that GazeRevealer achieves 77.43% detection accuracy for a single key number and 83.33% inference rate for the 6-digit password in the ideal case.

The rapid increase of multidimensional data (a.k.a. tensor) like videos brings new challenges for low-rank data modeling approaches such as dynamic data size, complex high-order relations, and multiplicity of low-rank structures. Resolving these challenges require a new tensor analysis method that can perform tensor data analysis online, which however is still absent. In this paper, we propose an Online Robust Low-rank Tensor Modeling (ORLTM) approach to address these challenges. ORLTM dynamically explores the high-order correlations across all tensor modes for low-rank structure modeling. To analyze mixture data from multiple subspaces, ORLTM introduces a new dictionary learning component. ORLTM processes data streamingly and thus requires quite low memory cost that is independent of data size. This makes ORLTM quite suitable for processing large-scale tensor data. Empirical studies have validated the effectiveness of the proposed method on both synthetic data and one practical task, i.e., video background subtraction. In addition, we provide theoretical analysis regarding computational complexity and memory cost, demonstrating the efficiency of ORLTM rigorously.

Amateur videos always contain focusing issues. A focusing mistake may produce out-of-focus blur, which seriously degrades the expressive force of the video. In this paper, we propose a patch-based method to remove the out-of-focus blur of a video and build an all-in-focus video. We assume that the out-of-focus blurry region in one frame will be clear in a portion of other frames; thus, the clear corresponding regions can be used to reconstruct the blurry one. We divide each video frame into a grid of patches and track each patch in the surrounding frames. We independently reconstruct each video frame by building a Markov random field model to identify the optimal target patches that are sharp, similar to the original patches, and are coherent with their neighboring patches within the overlapped regions. To recover an all-in-focus video, an iterative framework is utilized, in which the reconstructed video of each iteration is substituted in the next iteration. Finally, we employ the idea of a bilateral filter to temporally smooth the reconstructed video. The experimental results and the comparison with the previous works demonstrate the effectiveness of our method.

Chart classification is the foundation of chart analysis and document understanding. In this paper, we propose a novel framework to classify charts by combining convolutional networks and deep belief networks. In the framework, we firstly extract deep hidden features of charts, which are taken from the fully-connected layer of deep convolutional networks. We then utilize deep belief networks to predict the labels of the charts based on their deep hidden features. The convolutional networks are initialized using a large number of natural images and fine-tuned using the chart images to prevent overfitting. Compared with previous methods using primitive feature extraction, the deep features give our framework better scalability and stability. We collect a 5-class chart dataset with more than 5000 images and show that the proposed framework outperforms existing methods greatly.

In multi-hop wireless ad-hoc networks, packet delivery delay is one of the most important performance metrics. While a lot of research efforts have been spent on measuring and optimizing the end-to-end delay performance, there usually lack accurate and lightweight methods for decomposing the end-to-end delay into the per-hop delay for each packet. Knowledge on the per-hop per-packet delay can greatly improve the network visibility and facilitate network measurement and management. In this paper, we propose Domo, a passive, lightweight and accurate delay tomography approach to decomposing the packet end-to-end delay into each hop. The basic idea is to formulate the problem into a set of optimization problems by carefully considering the constraints among various timing quantities. At the network side, Domo attaches a small overhead to each packet for constructing constraints of the optimization problems. At the PC side, Domo employs semi-definite relaxation and several other methods to efficiently solve the optimization problems. We implement Domo and evaluate its performance extensively using large-scale simulations. Results show that Domo significantly outperforms two existing methods, nearly tripling the accuracy of the state-of-the-art.

Imitation is considered to be a kind of social learning that allows the transfer of information, actions, behaviours, etc. Whereas current robots are unable to perform as many tasks as human, it is a natural way for them to learn by imitations, just as human does. With the humanoid robots being more intelligent, the field of robot imitation has getting noticeable advance. In this paper, we focus on the pose imitation between a human and a humanoid robot and learning a similarity metric between human pose and robot pose. In contrast to recent approaches that capture human data using expensive motion captures or only imitate the upper body movements, our framework adopts a Kinect instead and can deal with complex, whole body motions by keeping both single pose balance and pose sequence balance. Meanwhile, different from previous work that employs subjective evaluation, we propose a pose similarity metric based on the shared structure of the motion spaces of human and robot. The qualitative and quantitative experimental results demonstrate a satisfactory imitation performance and indicate that the proposed pose similarity metric is discriminative.

Bulk data dissemination is a basic building block for enabling software update and reprogramming in wireless sensor networks. The recent structure based approach looks promising for efficient dissemination since it facilitates transmission and sleep scheduling. However, a number of limitations exist in existing structured protocols. In this paper, we propose a correlated core based solution for efficient bulk data dissemination in wireless sensor networks. We propose an efficient backbone node selection algorithm to construct the core structure by exploiting link correlation. We also design a novel negotiation mechanism which greatly reduces the control message overhead as compared to the existing structured protocols. We conduct both simulation and testbed experiments, and the results show that our proposed solution outperforms the state-of-the-art in terms of both the number of transmissions and the completion time.

Understanding the characteristics of MapReduce workloads in a Hadoop cluster is the key to making optimal configuration decisions and improving the system efficiency and throughput. However, workload analysis on a Hadoop cluster, particularly in a large-scale e-commerce production environment, has not been well studied yet. In this paper, we performed a comprehensive workload analysis using the trace collected from a 2000-node Hadoop cluster at Taobao, which is the biggest online e-commerce enterprise in Asia, ranked 10th in the world as reported by Alexa. The results of the workload analysis are representative and generally consistent with the data warehouses for e-commerce web sites, which can help researchers and engineers understand the workload characteristics of Hadoop in their production environments. Based on the observations and implications derived from the trace, we designed a workload generator Ankus, to expedite the performance evaluation and debugging of new mechanisms. Ankus supports synthesizing an e-commerce style MapReduce workload at a low cost. Furthermore, we proposed and implemented a job scheduling algorithm, Fair4S , which is designed to be biased towards small jobs. Small jobs account for the majority of the workload, and most of them require instant and interactive responses, which is an important phenomenon at production Hadoop systems. The inefficiency of Hadoop fair scheduler for handling small jobs motivates us to design the Fair4S, which introduces pool weights and extends job priorities to guarantee the rapid responses for small jobs. Experimental evaluation verified that the Fair4S accelerates the average waiting times of small jobs by a factor of 7 compared with the fair scheduler.

Previous packet length optimizations for sensor networks often employ a fixed optimal length scheme, while in this study we present DPLC, a Dynamic Packet Length Control scheme. To make DPLC more efficient in terms of channel utilization, we incorporate a lightweight and accurate link estimation method. We further provide two easy-to-use services, i.e., small message aggregation and large message fragmentation, to facilitate upper-layer application programming. The implementation of DPLC based on TinyOS 2.1 is lightweight, with respect to computation, memory, and header overhead. Our experiments using a real indoor testbed running CTP show that DPLC achieves the best performance compared with previous works.

Reprogramming over the air is important for maintaining a wireless sensor network. Traditional reprogramming approaches assume always-on link layers. Given the energy limitation of sensor nodes, always-on link layers are often not desired for most sensor network applications. In this paper, we propose ROLP, a novel reprogramming protocol built on a widely used low power link layer in wireless sensor networks. ROLP employs an efficient control packets self-suppression scheme for reliable data transmission. ROLP also employs an adaptively falling asleep scheme based on the neighbor information to reduce the energy consumption. We implement ROLP based on TinyOS and evaluate its performance in two different indoor networks. Compared with the standard reprogramming protocol in TinyOS, ROLP is able to reduce the radio-on time by 57.6% and 39.0% in the two networks. Since radio operations cost most of the energy, these reductions save significant amount of energy and prolong the lifetime of a wireless sensor network.

SIMPLE (Semi-Implicit Method for Pressure-Linked Equations) algorithm is important in the simulation of steady flows. As the traditional 3-D SIMPLE algorithm is time-consuming, we propose a parallel SIMPLE algorithm based on a novel tiling strategy -- alternate tiling, through replacing the original linear system and reordering the iteration space tiles. The novelty of our parallel algorithm lies in the introduction of the sequence of iteration space tiles as the sequence of execution, the time skewing technique to partition the iteration space, update operations of the grids from two directions alternately, and the improvement of the data locality. The effectiveness of the parallel algorithm and serial model of finite difference stencil algorithm are validated. Numerical experiments on distributed clusters show that the cache misses and the cost of communication and synchronization are reduced by reordering the tiles of iteration space, and the parallel SIMPLE algorithm based on alternate tiling has a good data locality and parallel efficiency in the three-dimensional incompressible pipe flow project.

Nodes of wireless sensor networks (WSN) may fail easily due to the lack of energy or disaster scenarios. Such failures can severely reduce the persistence and collection efficiency of sensed data. Network coding technology can be employed to enhance the data persistence of wireless sensor network, but it may cause serious "cliff effect" in decoding process. In this paper, the influence of prioritized coding degree distribution strategy on cliff effect is observed, and a distributed storage algorithm PLTD-Alpha is proposed. With PLTD-Alpha, the data in sensor network nodes present a trend that their degree distribution increase along with the degree level predefined, and the persistent data packets can be submitted to the sink node according to its degree in order. Experiment results show that PLTD-Alpha can greatly improve the data collection and decoding efficiency of sensor network while data persistence is not notably affected.

Wireless sensor network (WSN) is a typical distributed storage system, and network coding technology is developed to enhance the data persistence of WSN. However, the traditional distributed coding strategy may cause serious "cliff effect" in the decoding process, that is to say, few source data can be recovered before sufficient encoded packets are received. Moreover, nodes may fail due to the lack of energy or the influence of the switching of external environment, such as a disaster scenario. Such failures may concentrate in a small region or distribute in the whole deployment area which can severely reduced the decoding efficiency of the persistent data in WSN. In this paper, we propose the PLTCDS (prioritized LT codes based distributed storage) algorithm to improve the data decoding efficiency when the data persistence is assured. The main idea of PLTCDS is that the predefined node broadcasts a beacon to stimulate the nodes to form the network with degree distribution priority. To ensure the effectiveness of storage nodes, PLTCDS introduces a class of cumulative counter scheme to avoid empty storage. Also we discuss about the idea of another type of PLTCDS. Experimental results show that PLTCDS algorithm can enhance the network data collection performance and reduce the influence of "cliff effect".

In order to protect the privacy and the security of the tag, we propose a mutual authentication protocol for low-cost RFID system. The protocol ensures that the reader (or server) and the tag authenticate each other, provides the privacy and security properties for low-cost RFID system and also has good performance. The protocol resists almost all attacks except cloning attack, and achieves backward secrecy without any assumption and forward secrecy under an assumption. We also show that it requires less storage than other related RFID protocols.

This paper investigates event localization in wireless sensor networks. We improve the SNAP (Subtract on Negative Add on Positive) localization algorithm and propose the MSNAP (Modified Subtract on Negative Add on Positive) localization algorithm with higher localization accuracy and better performance of fault tolerance. First, every sensor node obverses the event signal and compares its observed reading with a threshold. If the reading is above the threshold, the node will send it to the sink station. Otherwise, it remains silent. Based on the observed readings which the nodes report, the sink station constructs the likelihood matrix by simply adding ± 1 contributions in the area around the nodes, whose maximum value points to the event location. Compared with the SNAP algorithm, when constructing the likelihood matrix, MSNAP dynamically adjusts the size of estimated region depending on the observed readings the nodes reported. Experimental results show that the algorithm effectively improves the localization accuracy and fault tolerance.

Growth Codes (GC) enhances the data persistence in dense sensor networks. However, GC exchanges data with neighbors in a completely random way, which may lead to uneven sensor data distribution in sparse sensor network. This significantly reduces the efficiency of GC data acquisition and fault-tolerance in sparse sensor network with less connectivity. In this paper, we propose Regulative Growth Codes (RGC) which use a random sequencing policy in data exchange operation instead of the completely random policy used in GC, and introduce the self-detection mechanism to reduce the redundant exchange. Simulation results show that the performance of RGC is better than GC in sparse sensor networks.

The performance evaluation of wireless sensor network based on passive monitoring restricted by two respects: first, passive monitoring trace is incomplete, because monitors can not capture every transmission in the network. Second, the information directly extracted from trace is insufficient. Performance evaluation always needs some implicit information, e.g., packet reception. To solving the problems above, we use two processing steps to construct an enhanced trace of network. First, an online merging procedure combines the incomplete traces of various monitors into a single more complete trace. Next, an inference procedure based on finite state machine reconstructs packets that were not captured by any monitor and determines whether a packet was received by its destination. Merging and inference procedure are realized in CTP Network which is incorporated with TinyOS. The evaluation is performed on simulation. The experiment results show that, if the trace contains more than 70% of the total packets, the engine can infer 20%-25% more loss packets and 90% packets' reception.

With the scale of computing system increases, power consumption has become the major challenge to system performance, reliability and IT management costs. Specifically, system performance and reliability, described by various Quality of Service(QoS) metrics, cannot be guaranteed if the objective is to minimize the total power consumption solely, despite of the violations of QoS. Various methods have been developed to control power consumption to avoid system failures and thermal emergencies through coarse-grained designs. However, the existing methods can be improved and more power can be saved if fine-grained job level adaptation is integrated into them. In this paper a feedback control based power aware job scheduling algorithm is proposed to minimize power consumption in computing system and to provide QoS guarantees. In the proposed algorithm, jobs are scheduled according to the realtime and historical power consumption as well as the QoS requirements. Simulations and experiments on real multi core computing system show that the power potential of the system can be deeply explored while still providing QoS guarantees and the performance degradation is acceptable. The experiment results also show that fine-grained job-level power aware scheduling can achieve better power/performance balancing between multiple processors or cores than coarse-grained methods.

With the growth of hardware and software resources, It becomes a challenge that how to ensure the Quality of Service (QoS) of resource scheduling in multiple virtual machine system. In order to solve this problem, the authors propose a control mechanism about QoS for resource scheduling. In the control mechanism, we first propose a theoretical model. Then, we present an elicitation algorithm to resolve the optimal model. In order to justify the feasibility and availability of this control mechanism about QoS for resource scheduling in multiple virtual machine system, a series of experiments have been done. The results show that it is feasible to schedule the system resources and control the QoS of system resources for tasks in multiple virtual machine system.

How to meet the QoS of the applications and improve resource utilization is an important problem in virtualized environment. In this paper, we propose a utility based resource allocation policy with QoS constrained in virtualized environment. Firstly, we build a model reflecting the mapping relation between performance metrics and resource allocation through Web server benchmarking experiments. Then, we devise a utility function as objective to optimizing the total utility and achieving a reasonable resource allocation. Finally, based on the model and the requirements of the applications QoS, we present an optimized policy for resource allocation in virtualized system. The simulation results show that the utility-based policy is effective to allocate resource from user perspective while improving the resource utilization.

In order to improve the resource utilization of virtual machine and control the resource allocation online effectively, in this paper, we present a grey prediction control model used for dynamic resource allocation in virtual machine as workloads changing. First, we forecast the allocation of virtualized resources by the grey control model. We also adjust the boundary conditions of grey prediction model to make the prediction more accurately. Then, the control theory is used to feedback control resource utilization to obtain desired resource utilization levels by regulating the value of allocation of virtualized resources automatically. Our experimental results show the grey control model is effective in the virtualized resource allocation. The control model and algorithm can be applied to other resource allocation.

In traditional grid, grid user can not validate whether grid service registered in grid platform can execute correctly or not because the system does not provide the measurement mechanism for grid service. In order to solve this problem, we propose a series of strategies and methods for trusted grid. These strategies and methods include: an access control policy for reference database, an integrity attestation for reference datasheet, a construction method of reference datasheet, a trusted storage algorithm of reference datasheet and a report algorithm of reference datasheet. All these constitute the measurement mechanism of service for trusted grid. Some experiments are done in order to validate the feasibility and the availability of this mechanism. The results show that it is feasible and efficient to validate whether grid service registered in trusted grid platform can execute correctly or not.

Grids provide infrastructures and solutions to solve large scale cooperative problems such as large scale distributed virtual environment simulation, multi-institutional scientific computing and data analysis, etc. In this paper, the key techniques for Grid computing based large scale Distributed Cooperative Virtual Environment Simulation (GDCVES) are discussed and a hierarchical architecture of GDCVES is proposed. The solutions of GDCVES, such as resource management, massive data management, security aware task scheduling, and fault-tolerance are also discussed. To evaluate the feasibility and scalability of GDCVES, a prototype was implemented and the simulation workflow framework is also analyzed.

Semantic-based image retrieval using low-level visual features is a challenging and important issue in content-based image retrieval. In this paper, we cast the image retrieval issue in a Bayesian framework and AAM (the active appearance model). Specifically, we propose an approach for complex semantic-based image retrieval, for example selecting the grassland images including horses. That is, the approach is used for selecting images including specific scene and model. In the approach, we integrate low-level features and spatial distribution into Bayesian frame. The approach uses Bayesian framework to select the images including the scene (forest, grassland), and uses AAM to select the images including the specific model (horse). Experimental results indicate that our approach is effective in complex semantic-based image retrieval and provides a sound retrieval performance.

The paper proposes a parameter clustering method using fuzzy clustering analysis to decrease the parameter size efficiently and improve the robustness of parameter training in large vocabulary continuous speech recognition systems. Based on the structure of a phonetic decision tree, leaf nodes are used for Gaussian fuzzy clustering and the root node or shallow leaf nodes are used for covariance fuzzy sharing. Experimental results show that, when the number of Gaussians is reduced by 50%, recognition accuracy only decreases by 0.55% compared to the baseline. By combining covariance fuzzy sharing, a significant performance increasing is achieved over the conventional system with approximately the same parameter size.

In order to improve the accuracy of the vertical handoff decision for radio heterogeneous network, this paper proposes an adaptive multi-criteria vertical handoff (AMVHO) decision algorithm. This algorithm uses a fuzzy inference system (FIS) and a modified Elman neural network (MENN). The FIS adopts crucial criteria of the vertical handoff as the input variables and makes handoff decision based on the defined rule base. The MENN helps to do the prediction for number-of-users of the after-handoff network, which is a pivotal variable of the FIS. Simulation results show that compared with the conventional method, the AMVHO decision algorithm achieves better performance in guaranteeing the quality of service (QoS) of the after-handoff communication.

Phonetic decision tree based state tying has been widely used in most large vocabulary continuous speech recognition (LVCSR) systems. However, in most cases, the samples of different leaf nodes are very unbalanced, which may affect the recognition performance. In This work, node merging techniques are proposed to alleviate the problem and further decrease the number of senones. On the other hand, in order to lessen the impact of rare triphones on the quality of the decision tree based state tying and improve the accuracy of every final senone, two methods of dealing with rare triphones are added to hidden Markov model (HMM) acoustic modeling before state tying. Experimental results show that these methods greatly improve the robustness of the decision tree and can achieve better performance with even fewer parameters.

The ability to undo operations is an indispensable feature of single user editing systems, but supporting group undo in real-time collaborative editing systems is still a difficult problem. In this paper, we propose an undo framework and algorithms to achieve group undo in image-based collaborative graphics editing systems. The basic idea is to interpret an undo command as a concurrent inverse operation by means of image operation transformation algorithm, so that an operation is always undoable under its current context. Through exploiting relations among operations, space cost for operation preservation is greatly reduced. The global undo, local undo and selective undo mode are supported in our solution. The undo algorithms are also applicable in single-user applications. The algorithms are implemented in CoDesign-a multi-level collaborative graphics designing system, which aims at supporting both object-based and image-based collaborative pattern design.