Speaker discrimination based on fuzzy fusion and feature reduction techniques

In this paper, we propose a research work on speaker discrimination using a multi-classifier fusion with focus on feature reduction effects. Speaker discrimination consists in the automatic distinction between two speakers using the vocal characteristics of their speeches. A number of features are extracted using Mel Frequency Spectral Coefficients and then reduced using Relative Speaker Characteristic (RSC) along with the Principal Components Analysis (PCA). Several classification methods are implemented to ensure the discrimination task. Since different classifiers are employed, two fusion algorithms at the decision level, referred to as Weighted Fusion and Fuzzy Fusion, are proposed to boost the classification performances. These algorithms are based on the weighting of the different classifiers outputs. Furthermore, the effects of speaker gender and feature reduction on the speaker discrimination task have been examined too. The evaluation of our approaches was conducted on a subset of Hub-4 Broadcast-News. The experimental results have shown that the speaker discrimination accuracy is improved by 5–15% using the (RSC–PCA) feature reduction. In addition, the proposed fusion methods recorded an improvement of about 10% compared to the individual scores of the classifiers. Finally, we noticed that the gender has an important impact on the discrimination performances.     

Locally adaptive <Emphasis Type="Italic">k</Emphasis> parameter selection for nearest neighbor classifier: one nearest cluster

The k nearest neighbors (k-NN) classification technique has a worldly wide fame due to its simplicity, effectiveness, and robustness. As a lazy learner, k-NN is a versatile algorithm and is used in many fields. In this classifier, the k parameter is generally chosen by the user, and the optimal k value is found by experiments. The chosen constant k value is used during the whole classification phase. The same k value used for each test sample can decrease the overall prediction performance. The optimal k value for each test sample should vary from others in order to have more accurate predictions. In this study, a dynamic k value selection method for each instance is proposed. This improved classification method employs a simple clustering procedure. In the experiments, more accurate results are found. The reasons of success have also been understood and presented.     

Fast moving pedestrian detection based on motion segmentation and new motion features

The detection of moving pedestrians is of major importance for intelligent vehicles, since information about such persons and their tracks should be incorporated into reliable collision avoidance algorithms. In this paper, we propose a new approach to detect moving pedestrians aided by motion analysis. Our main contribution is to use motion information in two ways: on the one hand we localize blobs of moving objects for regions of interest (ROIs) selection by segmentation of an optical flow field in a pre-processing step, so as to significantly reduce the number of detection windows needed to be evaluated by a subsequent people classifier, resulting in a fast method suitable for real-time systems. On the other hand we designed a novel kind of features called Motion Self Difference (MSD) features as a complement to single image appearance features, e. g. Histograms of Oriented Gradients (HOG), to improve distinctness and thus classifier performance. Furthermore, we integrate our novel features in a two-layer classification scheme combining a HOG+Support Vector Machines (SVM) and a MSD+SVM detector. Experimental results on the Daimler mono moving pedestrian detection benchmark show that our approach obtains a log-average miss rate of 36 % in the FPPI range [10^?2,10⁰], which is a clear improvement with respect to the naive HOG+SVM approach and better than several other state-of-the-art detectors. Moreover, our approach also reduces runtime per frame by an order of magnitude.     

Evaluation of k-Nearest Neighbor classifier performance for direct marketing

Text data mining is a process of exploratory data analysis. Classification maps data into predefined groups or classes. It is often referred to as supervised learning because the classes are determined before examining the data. This paper describes the proposed k-Nearest Neighbor classifier that performs comparative cross-validation for the existing k-Nearest Neighbor classifier. The feasibility and the benefits of the proposed approach are demonstrated by means of data mining problem: direct marketing. Direct marketing has become an important application field of data mining. Comparative cross-validation involves estimation of accuracy by either stratified k-fold cross-validation or equivalent repeated random subsampling. While the proposed method may have a high bias; its performance (accuracy estimation in our case) may be poor due to a high variance. Thus the accuracy with the proposed k-Nearest Neighbor classifier was less than that with the existing k-Nearest Neighbor classifier, and the smaller the improvement in runtime the larger the improvement in precision and recall. In our proposed method we have determined the classification accuracy and prediction accuracy where the prediction accuracy is comparatively high.     

Motor Imagery Electroencephalograph Classification Based on Optimized Support Vector Machine by Magnetic Bacteria Optimization Algorithm

In this paper, an optimized support vector machine (SVM) based on a new bio-inspired method called magnetic bacteria optimization algorithm method is proposed to construct a high performance classifier for motor imagery electroencephalograph based brain–computer interface (BCI). Butterworth band-pass filter and artifact removal technique are combined to extract the feature of frequency band of the ERD/ERS. Common spatial pattern is used to extract the feature vector which are put into the classifier later. The optimization mechanism involves kernel parameters setting in the SVM training procedure, which significantly influences the classification accuracy. Our novel approach aims to optimize the penalty factor parameter C and kernel parameter g of the SVM. The experimental results on the BCI Competition IV dataset II-a clearly present the effectiveness of the proposed method outperforming other competing methods in the literature such as genetic algorithm, particle swarm algorithm, artificial bee colony, biogeography based optimization.     

Profiling drivers based on driver dependent vehicle driving features

This work addresses the problem of profiling drivers based on their driving features. A purpose-built hardware integrated with a software tool is used to record data from multiple drivers. The recorded data is then profiled using clustering techniques. k-means has been used for clustering and the results are counterchecked with Fuzzy c-means (FCM) and Model Based Clustering (MBC). Based on the results of clustering, a classifier, i.e., an Artificial Neural Network (ANN) is trained to classify a driver during driving in one of the four discovered clusters (profiles). The performance of ANN is compared with that of a Support Vector Machine (SVM). Comparison of the clustering techniques shows that different subsets of the recorded dataset with a diverse combination of attributes provide approximately the same number of profiles, i.e., four. Analysis of features shows that average speed, maximum speed, number of times brakes were applied, and number of times horn was used provide the information regarding drivers’ driving behavior, which is useful for clustering. Both one versus one (SVM) and one versus rest (SVM) method for classification have been applied. Average accuracy and average mean square error achieved in the case of ANN was 84.2 % and 0.05 respectively. Whereas the average performance for SVM was 47 %, the maximum performance was 86 % using RBF kernel. The proposed system can be used in modern vehicles for early warning system, based on drivers’ driving features, to avoid accidents.     

A novel ship classification approach for high resolution SAR images based on the BDA-KELM classification model

Ship classification based on synthetic aperture radar (SAR) images is a crucial component in maritime surveillance. In this article, the feature selection and the classifier design, as two key essential factors for traditional ship classification, are jointed together, and a novel ship classification model combining kernel extreme learning machine (KELM) and dragonfly algorithm in binary space (BDA), named BDA-KELM, is proposed which conducts the automatic feature selection and searches for optimal parameter sets (including the kernel parameter and the penalty factor) for classifier at the same time. Finally, a series of ship classification experiments are carried out based on high resolution TerraSAR-X SAR imagery. Other four widely used classification models, namely k-Nearest Neighbour (k-NN), Bayes, Back Propagation neural network (BP neural network), Support Vector Machine (SVM), are also tested on the same dataset. The experimental results shows that the proposed model can achieve a better classification performance than these four widely used models with an classification accuracy as high as 97% and encouraging results of other three multi-class classification evaluation metrics.     

A Feature-Based Serial Approach to Classifier Combination

: A new approach to the serial multi-stage combination of classifiers is proposed. Each classifier in the sequence uses a smaller subset of features than the subsequent classifier. The classification provided by a classifier is rejected only if its decision is below a predefined confidence level. The approach is tested on a two-stage combination of k-Nearest Neighbour classifiers. The features to be used by the first classifier in the combination are selected by two stand-alone algorithms (Relief and Info-Fuzzy Network, or IFN) and a hybrid method, called ‘IFN + Relief’. The feature-based approach is shown empirically to provide a substantial decrease in the computational complexity, while maintaining the accuracy level of a single-stage classifier or even improving it. Received: 24 November 2000, Received in revised form: 30 November 2001, Accepted: 05 June 2002 ID="A1" Correspondence and offprint requests to: M. Last, Department of Information Systems Engineering, Ben-Gurion University of the Negev, Beer-Sheva 84105, Israel. Email: mlast@bgumail.bgu.ac.il     

<Emphasis Type="Italic">υ</Emphasis>-Support vector machine based on discriminant sparse neighborhood preserving embedding

In this paper, we mainly focus on two issues (1) SVM is very sensitive to noise. (2) The solution of SVM does not take into consideration of the intrinsic structure and the discriminant information of the data. To address these two problems, we first propose an integration model to integrate both the local manifold structure and the local discriminant information into ?₁ graph embedding. Then we add the integration model into the objection function of υ-support vector machine. Therefore, a discriminant sparse neighborhood preserving embedding υ-support vector machine (υ-DSNPESVM) method is proposed. The theoretical analysis demonstrates that υ-DSNPESVM is a reasonable maximum margin classifier and can obtain a very lower generalization error upper bound by minimizing the integration model and the upper bound of margin error. Moreover, in the nonlinear case, we construct the kernel sparse representation-based ?₁ graph for υ-DSNPESVM, which is more conducive to improve the classification accuracy than ?₁ graph constructed in the original space. Experimental results on real datasets show the effectiveness of the proposed υ-DSNPESVM method.     

Recognition of driving postures by multiwavelet transform and multilayer perceptron classifier

To develop Human-centric Driver Assistance Systems (HDAS) for automatic understanding and charactering of driver behaviors, an efficient feature extraction of driving postures based on Geronimo–Hardin–Massopust (GHM) multiwavelet transform is proposed, and Multilayer Perceptron (MLP) classifiers with three layers are then exploited in order to recognize four pre-defined classes of driving postures. With features extracted from a driving posture dataset created at Southeast University (SEU), the holdout and cross-validation experiments on driving posture classification are conducted by MLP classifiers, compared with the Intersection Kernel Support Vector Machines (IKSVMs), the k-Nearest Neighbor (kNN) classifier and the Parzen classifier. The experimental results show that feature extraction based on GHM multwavelet transform and MLP classifier, using softmax activation function in the output layer and hyperbolic tangent activation function in the hidden layer, offer the best classification performance compared to IKSVMs, kNN and Parzen classifiers. The experimental results also show that talking on a cellular phone is the most difficult one to classify among four predefined classes, which are 83.01% and 84.04% in the holdout and cross-validation experiments respectively. These results show the effectiveness of the feature extraction approach using GHM multiwavelet transform and MLP classifier in automatically understanding and characterizing driver behaviors towards Human-centric Driver Assistance Systems (HDAS).     

Exploiting multi-channels deep convolutional neural networks for multivariate time series classification

Time series classification is related to many different domains, such as health informatics, finance, and bioinformatics. Due to its broad applications, researchers have developed many algorithms for this kind of tasks, e.g., multivariate time series classification. Among the classification algorithms, k-nearest neighbor (k-NN) classification (particularly 1-NN) combined with dynamic time warping (DTW) achieves the state of the art performance. The deficiency is that when the data set grows large, the time consumption of 1-NN with DTWwill be very expensive. In contrast to 1-NN with DTW, it is more efficient but less effective for feature-based classification methods since their performance usually depends on the quality of hand-crafted features. In this paper, we aim to improve the performance of traditional feature-based approaches through the feature learning techniques. Specifically, we propose a novel deep learning framework, multi-channels deep convolutional neural networks (MC-DCNN), for multivariate time series classification. This model first learns features from individual univariate time series in each channel, and combines information from all channels as feature representation at the final layer. Then, the learnt features are applied into a multilayer perceptron (MLP) for classification. Finally, the extensive experiments on real-world data sets show that our model is not only more efficient than the state of the art but also competitive in accuracy. This study implies that feature learning is worth to be investigated for the problem of time series classification.     

Direction prediction for avoiding occlusion in visual surveillance

Occurrence of occlusion while providing visual surveillance leads to anarchy as the track of the subject under motion may be lost. This often results into the failure of the surveillance system. The approach of predicting motion of moving subjects and hence the chances of their mutual occlusion gives an upper hand to surveillance system to take in-time necessary action towards mitigation of loss of track during dynamic occlusion. Direction of motion of a moving subject plays a major role while studying its motion. Direction along with the velocity of a subject in a 3D plane completely describes the motion of any subject. This article proposes a model‘-based approach for direction prediction of a moving subject in a 3D global plane as acquired in a 2D camera plane. The proposed approach uses the eight discrete directions of motion as proposed in and models different directions. The proposed direction prediction method is experimentally verified with six different classifiers, i.e. regression analysis, simple logistic regression, MLP, k-NN, SVM and Bays classifier over existing as well as self-acquired databases. The initial simulation results are motivating as the overall accuracies achieved through different classifiers are of the range of 87–94 \(\%\), which advocates the suitability of the said approach.     

Multiscale Local Gabor Phase Quantization for image forgery detection

Image Forgery is a field that has attracted the attention of a significant number of researchers in the recent years. The widespread popularity of imagery applications and the advent of powerful and inexpensive cameras are among the numerous reasons that have contributed to this upward spike in the reach of image manipulation. A considerable number of features – including numerous texture features – have been proposed by various researchers for identifying image forgery. However, detecting forgery in images utilizing texture-based features have not been explored to its full potential – especially a thorough evaluation of the texture features have not been proposed. In this paper, features based on image textures are extracted and combined in a specific way to detect the presence of image forgery. First, the input image is converted to YCbCr color space to extract the chroma channels. Gabor Wavelets and Local Phase Quantization are subsequently applied to these channels to extract the texture features at different scales and orientations. These features are then optimized using Non-negative Matrix Factorization (NMF) and fed to a Support Vector Machine (SVM) classifier. This method leads to the classification of images with accuracies of 99.33%, 96.3%, 97.6%, 85%, and 96.36% for the CASIA v2.0, CASIA v1.0, CUISDE, IFS-TC and Unisa TIDE datasets respectively showcasing its ability to identify image forgeries under varying conditions. With CASIA v2.0, the detection accuracy outperforms the recent state-of-the-art methods, and with the other datasets, it gives a comparable performance with much reduced feature dimensions.     

Fast graph similarity search via hashing and its application on image retrieval

Similarity search in graph databases has been widely investigated. It is worthwhile to develop a fast algorithm to support similarity search in large-scale graph databases. In this paper, we investigate a k-NN (k-Nearest Neighbor) similarity search problem by locality sensitive hashing (LSH). We propose an innovative fast graph search algorithm named LSH-GSS, which first transforms complex graphs into vectorial representations based on prototypes in the database and later accelerates a query in Euclidean space by employing LSH. Because images can be represented as attributed graphs, we propose an approach to transform attributed graphs into n-dimensional vectors and apply LSH-GSS to execute further image retrieval. Experiments on three real graph datasets and two image datasets show that our methods are highly accurate and efficient.     

An all-pair quantum SVM approach for big data multiclass classification

In this paper, we discuss a quantum approach for the all-pair multiclass classification problem. In an all-pair approach, there is one binary classification problem for each pair of classes, and so there are k(k???1)/2 classifiers for a k-class classification problem. As compared to the classical multiclass support vector machine that can be implemented with polynomial run time complexity, our approach exhibits exponential speedup due to quantum computing. The quantum all-pair algorithm can also be used with other classification algorithms, and a speedup gain can be achieved as compared to their classical counterparts.     

A heuristic supervised Euclidean data difference dimension reduction for KNN classifier and its application to visual place classification

In this paper, we propose a novel supervised dimension reduction algorithm based on K-nearest neighbor (KNN) classifier. The proposed algorithm reduces the dimension of data in order to improve the accuracy of the KNN classification. This heuristic algorithm proposes independent dimensions which decrease Euclidean distance of a sample data and its K-nearest within-class neighbors and increase Euclidean distance of that sample and its M-nearest between-class neighbors. This algorithm is a linear dimension reduction algorithm which produces a mapping matrix for projecting data into low dimension. The dimension reduction step is followed by a KNN classifier. Therefore, it is applicable for high-dimensional multiclass classification. Experiments with artificial data such as Helix and Twin-peaks show ability of the algorithm for data visualization. This algorithm is compared with state-of-the-art algorithms in classification of eight different multiclass data sets from UCI collection. Simulation results have shown that the proposed algorithm outperforms the existing algorithms. Visual place classification is an important problem for intelligent mobile robots which not only deals with high-dimensional data but also has to solve a multiclass classification problem. A proper dimension reduction method is usually needed to decrease computation and memory complexity of algorithms in large environments. Therefore, our method is very well suited for this problem. We extract color histogram of omnidirectional camera images as primary features, reduce the features into a low-dimensional space and apply a KNN classifier. Results of experiments on five real data sets showed superiority of the proposed algorithm against others.     

Asymptotic assessments of CRC error probabilities in some telecommunication protocols

We study the value distributions for the control cyclic redundancy check (CRC) of length k, drawn at the data section of volume n. The behavior of CRC value distribution is examined at large n and fixed values of k (k = const, n → ∞). With the application of the character theory, we find the conditions of asymptomatic uniformity of the CRC distribution. The asymptomatic results can be applied during the assessment of errors of a series of protocols such as USB, X.25, HDLC, Bluetooth, Ethernet, etc.     

Advanced methods for the estimation of an unknown projection from a map

This article presents three new methods (M5, M6, M7) for the estimation of an unknown map projection and its parameters. Such an analysis is beneficial and interesting for historic, old, or current maps without information about the map projection; it could improve their georeference. The location similarity approach takes into account the residuals on the corresponding features; the minimum is found using the non-linear least squares. For the shape similarity approach, the minimized objective function ? takes into account the spatial distribution of the features, together with the shapes of the meridians, parallels and other 0D-2D elements. Due to the non-convexity and discontinuity, its global minimum is determined using the global optimization, represented by the differential evolution. The constant values of projection φ _k , λ _k , φ ₁, λ ₀, and map constants R,ΔX,ΔY, α (in relation to which the methods are invariant) are estimated. All methods are compared and the results are presented for the synthetic data as well as for 8 early maps from the Map Collection of the Charles University and the David Rumsay Map Collection. The proposed algorithms have been implemented in the new version of the detectproj software.