Adaptive modulation is an efficient method to increase the spectral efficiency of OFDM based high-speed wireless data transmission systems in multipath channel. Blind modulation classification schemes play an important role in adaptive modulation systems, eliminating the need for transmitting modulation information, thereby increasing spectral efficiency. In this paper, a novel blind modulation classification algorithm is derived from the finite alphabet property of information symbols and the equivalent parallel model of OFDM systems. The performances of the proposed algorithm and M2M4P algorithm [1] are tested and compared using Monte-Carlo simulations. It is found that, the novel algorithm yields performance better than that of M2M4P algorithm and with much less complexity.

A feature for classification of shallowly buried landmine-like objects using a ground penetrating radar (GPR) measurement system is proposed and its performance is evaluated. The feature for classification employed here is a time interval between two pulses reflected from top and bottom sides of landmine-like objects. First, we estimate a time resolution required to detect object thickness from GPR data, and check the actual time resolution through laboratory experiment. Next, we evaluate the classification performance using Monte Carlo simulations from dataset generated by a two-dimensional finite difference time domain (FDTD) method. The results show that good classification performance is achieved even for landmine-like objects buried at shallow depths under rough ground surfaces. Furthermore, we also estimate the effects of ground surface roughness, soil inhomogeneity, and target inclination on the classification performance.

A semi-supervised classification method is presented. A robust unsupervised spectral mapping method is extended to a semi-supervised situation. Our proposed algorithm is derived by linearization of this nonlinear semi-supervised mapping method. Experiments using the proposed method for some public benchmark data reveal that our method outperforms a supervised algorithm using the linear discriminant analysis for the iris and wine data and is also more accurate than a semi-supervised algorithm of the logistic GRF for the ionosphere dataset.

In this letter, a new segment-level speech/nonspeech classification method based on the Poisson polling technique is proposed. The proposed method makes two modifications from the baseline Poisson polling method to further improve the classification accuracy. One of them is to employ Poisson mixture models to more accurately represent various segmental patterns of the observed frequencies for frame-level input features. The other is the soft counting-based frequency estimation to improve the reliability of the observed frequencies. The effectiveness of the proposed method is confirmed by the experimental results showing the maximum error reduction of 39% compared to the segmentally accumulated log-likelihood ratio-based method.

The Erlang capacity of multi-service multi-access systems supporting several different radio access technologies was analyzed and compared according to two different operation methods: the separate and common operation methods, by simultaneously considering the link capacity limit per sector as well as channel element (CE) limit in a base station (BS). In a numerical example with GSM-like and WCDMA-like sub-systems, it is shown that we can get up to 60% Erlang capacity improvement through the common operation method using a near optimum so-called service-based user assignment scheme when there is no CE limit in BS. Even with the worst-case assignment scheme, we can still get about 15% capacity improvement over the separate operation method. However, a limited number of CEs in BS reduces the capacity gains of multi-service multi-access systems in both the common operation and separate operation. In order to fully extract the Erlang capacity of multi-service multi-access systems, an efficient method is needed in order to select a proper number of CE in BS while minimizing the equipment cost.

Several research fields have to deal with very large classification problems, e.g. handwritten character recognition and speech recognition. Many works have proposed methods to address problems with large number of samples, but few works have been done concerning problems with large numbers of classes. CombNET-II was one of the first methods proposed for such a kind of task. It consists of a sequential clustering VQ based gating network (stem network) and several Multilayer Perceptron (MLP) based expert classifiers (branch networks). With the objectives of increasing the classification accuracy and providing a more flexible model, this paper proposes a new model based on the CombNET-II structure, the CombNET-III. The new model, intended for, but not limited to, problems with large number of classes, replaces the branch networks MLP with multiclass Support Vector Machines (SVM). It also introduces a new probabilistic framework that outputs posterior class probabilities, enabling the model to be applied in different scenarios (e.g. together with Hidden Markov Models). These changes permit the use of a larger number of smaller clusters, which reduce the complexity of the final classifiers. Moreover, the use of binary SVM with probabilistic outputs and a probabilistic decoding scheme permit the use of a pairwise output encoding on the branch networks, which reduces the computational complexity of the training stage. The experimental results show that the proposed model outperforms both the previous model CombNET-II and a single multiclass SVM, while presenting considerably smaller complexity than the latter. It is also confirmed that CombNET-III classification accuracy scales better with the increasing number of clusters, in comparison with CombNET-II.

This letter proposes a fast k nearest neighbors search algorithm based on the wavelet transform. This technique exploits the important information of the approximation coefficients of the transform coefficient vector, from which we obtain two crucial inequalities that can be used to reject those vectors for which it is impossible to be k nearest neighbors. The computational complexity for searching for k nearest neighbors can be largely reduced. Experimental results on texture classification verify the effectiveness of our algorithm.

The minimization problem of an L2-sensitivity measure subject to L2-norm dynamic-range scaling constraints is formulated for a class of two-dimensional (2-D) state-space digital filters. First, the problem is converted into an unconstrained optimization problem by using linear-algebraic techniques. Next, the unconstrained optimization problem is solved by applying an efficient quasi-Newton algorithm with closed-form formula for gradient evaluation. The coordinate transformation matrix obtained is then used to synthesize the optimal 2-D state-space filter structure that minimizes the L2-sensitivity measure subject to L2-norm dynamic-range scaling constraints. Finally, a numerical example is presented to illustrate the utility of the proposed technique.

This paper describes a method for the classification of bank-notes. The algorithm has three stages, and classifies bank-notes with very low error rates and at high speeds. To achieve the very low error rates, the result of classification is checked in the final stage by using different features to those used in the first two. High-speed processing is mainly achieved by the hierarchical structure, which leads to low computational costs. In evaluation on 32,850 samples of US bank-notes, with the same number used for training, the algorithm classified all samples precisely with no error sample. We estimate that the worst error rate is 3.1E-9 for the classification statistically.

A new method for automatically building statistical shape models from a set of training examples and in particular from a class of hands. In this study, we utilise a novel approach to automatically recover the shape of hand outlines from a series of 2D training images. Automated landmark extraction is accomplished through the use of the self-organising model the growing neural gas (GNG) network, which is able to learn and preserve the topological relations of a given set of input patterns without requiring a priori knowledge of the structure of the input space. The GNG is compared to other self-organising networks such as Kohonen and Neural Gas (NG) maps and results are given for the training set of hand outlines, showing that the proposed method preserves accurate models.

The first realization of a class-F InGaP/GaAs HBT amplifier considering up to 7th-order higher harmonic frequencies, operating at 1.9-GHz band, is described. A total number of open-circuited stubs for higher harmonic frequency treatment is successfully reduced without changing a class-F load circuit condition, using a low-cost and low-loss resin (tan δ=0.0023) circuit board. In class-F amplifier design at microwave frequency ranges, not only increasing treated orders of higher harmonic frequencies for a class-F load circuit, but also decreasing parasitic capacitances of a transistor is important. Influence of a base-collector capacitance, Cbc, for power added efficiency, PAE, and collector efficiency, ηc, was investigated by using a two-dimensional device simulator and a harmonic balance simulator. Measured maximum PAE and ηc reached 74.2% and 76.6%, respectively, using a fabricated class-F InGaP/GaAs HBT amplifier with collector doping density of 21016 cm-3. In case circuit losses were de-embedded for the experimental results, PAE and ηc were estimated as 78.7% and 81.2%, respectively. These are very close to obtainable maximum PAE for the use of the InGaP/GaAs HBT.

A multi-stage approach -- which is fast, robust and easy to train -- for a face-detection system is proposed. Motivated by the work of Viola and Jones [1], this approach uses a cascade of classifiers to yield a coarse-to-fine strategy to reduce significantly detection time while maintaining a high detection rate. However, it is distinguished from previous work by two features. First, a new stage has been added to detect face candidate regions more quickly by using a larger window size and larger moving step size. Second, support vector machine (SVM) classifiers are used instead of AdaBoost classifiers in the last stage, and Haar wavelet features selected by the previous stage are reused for the SVM classifiers robustly and efficiently. By combining AdaBoost and SVM classifiers, the final system can achieve both fast and robust detection because most non-face patterns are rejected quickly in earlier layers, while only a small number of promising face patterns are classified robustly in later layers. The proposed multi-stage-based system has been shown to run faster than the original AdaBoost-based system while maintaining comparable accuracy.

Active Shape Model (ASM) has been shown to be a powerful tool to aid the interpretation of images, especially in face alignment. ASM local appearance model parameter estimation is based on the assumption that residuals between model fit and data have a Gaussian distribution. Moreover, to generate an allowable face shape, ASM truncates coefficients of shape principal components into the bounds determined by eigenvalues. In this paper, an algorithm of modeling local appearances, called AdaBoosted ASM, and a shape parameter optimization method are proposed. In the algorithm of modeling the local appearances, we describe our novel modeling method by using AdaBoosted histogram classifiers, in which the assumption of the Gaussian distribution is not necessary. In the shape parameter optimization, we describe that there is an inadequacy on controlling shape parameters in ASM, and our novel method on how to solve it. Experimental results demonstrate that the AdaBoosted histogram classifiers improve robustness of landmark displacement greatly, and the shape parameter optimization solves the inadequacy problem of ASM on shape constraint effectively.

We consider the classification problem to construct a classifier c:{0,1}n{0,1} from a given set of examples (training set), which (approximately) realizes the hidden oracle y:{0,1}n{0,1} describing the phenomenon under consideration. For this problem, a number of approaches are already known in computational learning theory; e.g., decision trees, support vector machines (SVM), and iteratively composed features (ICF). The last one, ICF, was proposed in our previous work (Haraguchi et al., (2004)). A feature, composed of a nonempty subset S of other features (including the original data attributes), is a Boolean function fS:{0,1}S{0,1} and is constructed according to the proposed rule. The ICF algorithm iterates generation and selection processes of features, and finally adopts one of the generated features as the classifier, where the generation process may be considered as embodying the idea of boosting, since new features are generated from the available features. In this paper, we generalize a feature to an extended Boolean function fS:{0,1,*}S{0,1,*} to allow partial knowledge, where * denotes the state of uncertainty. We then propose the algorithm ICF* to generate such generalized features. The selection process of ICF* is also different from that of ICF, in that features are selected so as to cover the entire training set. Our computational experiments indicate that ICF* is better than ICF in terms of both classification performance and computation time. Also, it is competitive with other representative learning algorithms such as decision trees and SVM.

This paper investigates the design curves of class E amplifier with nonlinear capacitance for any output Q and finite dc-feed inductance. The important results are; 1) the capacitance nonlinearity strongly affects the design parameters for low Q, 2) the value of dc-feed inductance is hardly affected by the capacitance nonlinearity, and 3) the input voltage is an important parameter to design class E amplifier with nonlinear capacitance. By carrying out PSpice simulations, we show that the simulated results agree with the desired ones quantitatively. It is expected that the design curves in this paper are useful guidelines for the design of class E amplifier with nonlinear capacitance.

This paper presents an approach to specification-based testing of concurrent programs with representative test sequences generated from Statecharts. Representative test sequences are a subset of all possible interleavings of concurrent events that define the behaviors of a concurrent program. Because a program's correctness may be determined by checking whether a program implemented all behaviors in its specification or not, the program can be regarded as being correct if it can supply an alternative execution that has the same effects as the program's behavior with each representative test sequence. Based on this observation, we employ each representative test sequence as a seed to generate an automaton that accepts its equivalent sequences to reveal the same behavior. In order to effectively test a concurrent program, the automaton such generated accepts all sequences equivalent to the representative test sequence and it is used to control test execution. This paper describes an automated process of generating automata from a Statecharts specification and shows how the proposed approach works on Statecharts specifications through some examples.

Shadow detection and removal is important to deal with traffic image sequences. Cast shadow of vehicle may lead to an inaccurate object feature extraction and erroneous scene analysis. Furthermore, separate vehicles can be connected through shadow. Both can confuse object recognition systems. In this paper, a robust method is proposed for detecting and removing active cast shadow in monocular color image sequences. Background subtraction method is used to extract moving blobs in color and gradient dimensions, and the YCrCb color space is adopted for detecting and removing the cast shadow. Even when shadows link different vehicles, it can detect the each vehicle figure using modified mask by shadow bar. Experimental results from town scenes show that proposed method is effective and the classification accuracy is sufficient for general vehicle type classification.

This paper introduces a network-aware video delivery framework where the quality-of-service (QoS) interaction between prioritized packet video and relative differentiated service (DiffServ) network is taken into account. With this framework, we propose a dynamic class mapping (DCM) scheme to allow video applications to cope with service degradation and class-based resource constraint in a time-varying network environment. In the proposed scheme, an explicit congestion notification (ECN)-based feedback mechanism is utilized to notify the status of network classes and the received service quality assessment to the end-host applications urgently. Based on the feedback information, DCM agent at ingress point can dynamically re-map each packet onto a network class in order to satisfy the desired QoS requirement. Simulation results verify the enhanced QoS performance of the streaming video application by comparing the static class-mapping and the class re-mapping based on loss-driven feedback.

Differentiated-services model has been prevailed as a scalable solution to provide quality of service over the Internet. Many researches have been focused on per hop behavior or a single domain behavior to enhance quality of service. Thus, there are still difficulties in providing the end-to-end guaranteed service when the path between sender and receiver includes multiple domains. Furthermore differentiated-services model mainly considers quality of service for traffic aggregates due to the scalability, and the quality of service state may be time varying according to the network conditions in the case of relative service model, which make the problem more challenging to guarantee the end-to-end quality-of-service. In this paper, we study class mapping mechanisms along the path to provide the end-to-end guaranteed quality of service with the minimum networking price over multiple differentiated-services domains. The proposed mechanism includes an effective implementation of relative differentiated-services model, quality of service advertising mechanism and class selecting mechanisms. Finally, the experimental results are provided to show the performance of the proposed algorithm.

Since sentences are the basic propositional units of text, knowing their themes should help in completing various tasks such as automatic summarization requiring the knowledge about the semantic content of text. Despite the importance of determining the theme of a sentence, however, few studies have investigated the problem of automatically assigning a theme to a sentence. In this paper, we examine the notion of sentence theme and propose an automatic scheme where head-driven patterns are used for theme assignment. We tested our scheme with sentences in encyclopedia articles and obtained a promising result of 98.96% in F-score for training data and 88.57% for testing data, which outperform the baseline using all but the head-driven patterns.