The H.264/AVC video coding standard uses an intra prediction mode with 4×4 and 16×16 blocks for luma and 8×8 blocks for chroma. This standard uses the rate distortion optimization (RDO) method to determine the best coding mode based on the compression performance and video quality. This method offers a large improvement in coding efficiency compared to other compression standards, but the computational complexity is greater due to the various intra prediction modes. This paper proposes a fast intra mode decision algorithm for real-time encoding of H.264/AVC based on the dominant edge direction (DED). The DED is extracted using pixel value summation and subtraction in the horizontal and vertical directions. By using the DED, three modes instead of nine are chosen for RDO calculation to decide on the best mode in the 4×4 luma block. For the 16×16 luma and the 8×8 chroma, only two modes are chosen instead of four. Experimental results show that the entire encoding time saving of the proposed algorithm is about 67% compared to the full intra search method with negligible loss of quality.

The rate and distortion of I_d-slice do not fit the globally linear relationship on a logarithmic scale. Lagrange multiplier selection methods based on the globally linear approximate relationship are neither efficient nor optimal for multi-view video coding (MVC). To improve the coding efficiency of MVC, a local curve fitting based Lagrange multiplier selection method is proposed in this paper, where Lagrange multipliers are selected according to the local slopes of the approximate curves. Experimental results showed that the proposed method improves the coding efficiency. Up to 2.5 dB gain was achieved at low bitrates.

Existing water hazard detection methods usually fail when the features of water surfaces are greatly changed by the surroundings, e.g., by a change in illumination. This paper proposes a novel algorithm to robustly detect different kinds of water hazards for autonomous navigation. Our algorithm combines traditional machine learning and image segmentation and uses only digital cameras, which are usually affordable, as the visual sensors. Active learning is used for automatically dealing with problems caused by the selection, labeling and classification of large numbers of training sets. Mean-shift based image segmentation is used to refine the final classification. Our experimental results show that our new algorithm can accurately detect not only ‘common’ water hazards, which usually have the features of both high brightness and low texture, but also ‘special’ water hazards that may have lots of ripples or low brightness.

An experimental machine vision apparatus was used to identify and extract recyclable plastic bottles out of a conveyor belt. Color images were taken with a commercially available Webcam, and the recognition was performed by our homemade software, based on the shape and dimensions of object images. The software was able to manage multiple bottles in a single image and was additionally extended to cases involving touching bottles. The identification was fulfilled by comparing the set of measured features with an existing database and meanwhile integrating various recognition techniques such as minimum distance in the feature space, self-organized maps, and neural networks. The recognition system was tested on a set of 50 different bottles and provided so far an accuracy of about 97% on bottle identification. The extraction of the bottles was performed by means of a pneumatic arm, which was activated according to the plastic type; polyethylene-terephthalate (PET) bottles were left on the conveyor belt, while non-PET bottles were extracted. The software was designed to provide the best compromise between reliability and speed for real-time applications in view of the commercialization of the system at existing recycling plants.

Previously we have designed and implemented new image browsing facilities to support effective offline image contents on mobile devices with limited capabilities: low bandwidth, small display, and slow processing. In this letter, we fulfill the automatic production of cartoon contents fitting small-screen display, and introduce a clustering method useful for various types of cartoon images as a prerequisite stage for preserving semantic meaning. The usage of neural networks is to properly cut the various forms of pages. Texture information that is useful for grayscale image segmentation gives us a good clue for page layout analysis using the multilayer perceptron (MLP) based x-y recursive algorithm. We also automatically frame the segment MLP using agglomerative segmentation. Our experimental results show that the combined approaches yield good results of segmentation for several cartoons.

Some two-microphone noise reduction techniques that work in the frequency domain exploit coherence function between two noisy signals. They have shown good results when noise signals on two sensors are uncorrelated, but their performance decreases with correlated noises. Coherence based methods can be improved when the cross power spectral density (CPSD) of correlated noise signals is available. In this paper, we propose a new method for estimation of the CPSD of the noise, which is based on the minimum tracking technique. Despite the fact that the proposed estimator does not need to implement a voice activity detector (VAD), its performance is comparable to a CPSD estimator that uses an ideal VAD.

We present a novel model for recognizing long-term complex activities involving multiple persons. The proposed model, named ‘decomposed hidden Markov model’ (DHMM), combines spatial decomposition and hierarchical abstraction to capture multi-modal, long-term dependent and multi-scale characteristics of activities. Decomposition in space and time offers conceptual advantages of compaction and clarity, and greatly reduces the size of state space as well as the number of parameters. DHMMs are efficient even when the number of persons is variable. We also introduce an efficient approximation algorithm for inference and parameter estimation. Experiments on multi-person activities and multi-modal individual activities demonstrate that DHMMs are more efficient and reliable than familiar models, such as coupled HMMs, hierarchical HMMs, and multi-observation HMMs.

We propose an adaptive fractional window increasing algorithm (AFW) to improve the performance of the fractional window increment (FeW) in (Nahm et al., 2005). AFW fully utilizes the bandwidth when the network is idle, and limits the operating window when the network is congested. We evaluate AFW and compare the total throughput of AFW with that of FeW in different scenarios over chain, grid, random topologies and with hybrid traffics. Extensive simulation through ns2 shows that AFW obtains 5% higher throughput than FeW, whose throughput is significantly higher than that of TCP-Newreno, with limited modifications.

Traditional 802.11 power saving mechanism (PSM) treats multicast and broadcast traffic equally, and suffers significant performance degradation with multicast background traffic. This paper proposes an enhanced PSM that effectively differentiates multicast streams. It re-arranges the virtual bitmap of the traffic indication map (TIM) to carry traffic status for multicast groups and introduces a concept of sequential transmission of multi-addressed data to facilitate differentiation among multicast groups. Our analysis shows that the enhanced PSM can effectively save power in mixed traffic environments.

We present a new normal basis multiplication scheme using a multiplexer-based algorithm. In this algorithm, the proposed multiplier processes in parallel and has a multiplexer-based structure that uses MUX and XOR gates instead of AND and XOR gates. We show that our multiplier for type-1 and type-2 normal bases saves about 8% and 16%, respectively, in space complexity as compared to existing normal basis multipliers. Finally, the proposed architecture has regular and modular configurations and is well suited to VLSI implementations.

With the rapid growth of the Web, the volume of information on the Web is increasing exponentially. However, information on the current Web is only understandable to humans, and this makes precise information retrieval difficult. To solve this problem, the Semantic Web was proposed. We must use ontology languages that can assign data the semantics for realizing the Semantic Web. One of the representative ontology languages is the Web ontology language OWL, adopted as a recommendation by the World-Wide Web Consortium (W3C). OWL includes hierarchical structural information between classes or properties. Therefore, an efficient OWL storage model that considers a hierarchical structure for effective information retrieval on the Semantic Web is required. In this paper we suggest an XPath-based OWL storage (XPOS) model, which includes hierarchical information between classes or properties in XPath form, and enables intuitive and effective information retrieval. Also, we show the comparative evaluation results for the performance of the XPOS model, Sesame, and the XML file system-based storage (XFSS) model, in terms of query processing and ontology updating.

This paper deals with the statistical modeling of latent topic hierarchies in text corpora. The height of the topic tree is assumed as fixed, while the number of topics on each level as unknown a priori and to be inferred from data. Taking a nonparametric Bayesian approach to this problem, we propose a new probabilistic generative model based on the nested hierarchical Dirichlet process (nHDP) and present a Markov chain Monte Carlo sampling algorithm for the inference of the topic tree structure as well as the word distribution of each topic and topic distribution of each document. Our theoretical analysis and experiment results show that this model can produce a more compact hierarchical topic structure and captures more fine-grained topic relationships compared to the hierarchical latent Dirichlet allocation model.

By considering the static voltage characteristic of the load, we propose a WAMS/SCADA mixed nonlinear method to estimate the voltage of unobservable buses caused by topology change or phasor measurement unit (PMU) malfunction in a power system. By modeling the load characteristic with data from SCADA, we employed the Gauss-Seidel method to solve the nonlinear equations and estimate the voltage of unobservable buses with the high precision voltages of neighboring buses measured by a PMU. Simulations were carried out on the IEEE 39-bus system, and the results show that this novel method can dynamically and accurately trace the variation of the voltage phasor of the unobservable buses.

This paper presents a new method based on an immune-tabu hybrid algorithm to solve the thermal unit commitment (TUC) problem in power plant optimization. The mathematical model of the TUC problem is established by analyzing the generating units in modern power plants. A novel immune-tabu hybrid algorithm is proposed to solve this complex problem. In the algorithm, the objective function of the TUC problem is considered as an antigen and the solutions are considered as antibodies, which are determined by the affinity computation. The code length of an antibody is shortened by encoding the continuous operating time, and the optimum searching speed is improved. Each feasible individual in the immune algorithm (IA) is used as the initial solution of the tabu search (TS) algorithm after certain generations of IA iteration. As examples, the proposed method has been applied to several thermal unit systems for a period of 24 h. The computation results demonstrate the good global optimum searching performance of the proposed immune-tabu hybrid algorithm. The presented algorithm can also be used to solve other optimization problems in fields such as the chemical industry and the power industry.

A novel control strategy for the load converter supplying the unbalanced AC load in a DC isolated distribution system is presented. The control algorithm results in balanced and sinusoidal load voltages under unbalanced AC loading. The unbalanced load is characterized in the d-q-0 rotating coordinate based on symmetrical sequence components. Also, the mathematical model of the load converter in both a-b-c and d-q-0 coordinates is derived by using the average large signal model. Then, two control strategies for the load converter are presented. The first one uses the conventional d-q-0 controller to ensure the voltage and current regulation. The second one is a newly proposed control strategy based on the decomposition of the voltage and current into instantaneous positive, negative, and zero sequences. These three sequences are controlled independently in their own reference frames as DC signals. The performance of the load converter using these two control strategies is compared. Simulation results show the validity and capability of the newly proposed control strategy.

To improve the handling performance of a steer-by-wire (SBW) vehicle, a series of control logics are proposed. Firstly, an algorithm for enhancing the maneuvering in steady-state cornering is presented. On this basis, two categories of control strategies are used to dynamically correct and compensate the transient state steering responses and vehicle behaviors. Simulator tests including subjective evaluations and virtual field tests are both conducted to make comprehensive investigations on the series of control logics. The subjective evaluations demonstrate that the SBW vehicle with a specifically selected value of steering sensitivity tends to be more desirable for driving than a conventional one in which a fixed steering ratio exists. The virtual field tests indicate that the control strategies for dynamical correction and compensation could effectively improve the handling performances of an SBW vehicle by reducing the work load of drivers, enhancing the track-holding performance, and improving steering response properties.

This study compares the ability of different robust regression estimators to detect and classify outliers. Well-known estimators with high breakdown points were compared using simulated data. Mean success rates (MSR) were computed and used as comparison criteria. The results showed that the least median of squares (LMS) and least trimmed squares (LTS) were the most successful methods for data that included leverage points, masking and swamping effects or critical and concentrated outliers. We recommend using LMS and LTS as diagnostic tools to classify outliers, because they remain robust even when applied to models that are heavily contaminated or that have a complicated structure of outliers.

The evolvable multiprocessor (EvoMP), as a novel multiprocessor system-on-chip (MPSoC) machine with evolvable task decomposition and scheduling, claims a major feature of low-cost and efficient fault tolerance. Non-centralized control and adaptive distribution of the program among the available processors are two major capabilities of this platform, which remarkably help to achieve an efficient fault tolerance scheme. This letter presents the operational as well as architectural details of this fault tolerance scheme. In this method, when a processor becomes faulty, it will be eliminated of contribution in program execution in remaining run-time. This method also utilizes dynamic rescheduling capability of the system to achieve the maximum possible efficiency after processor reduction. The results confirm the efficiency and remarkable advantages of the proposed approach over common redundancy based techniques in similar systems.