A Drude-critical points (D-CP) model for considering metal dispersion is newly incorporated into the frequency-dependent FDTD method using the simple trapezoidal recursive convolution (TRC) technique. Numerical accuracy is investigated through the analysis of pulse propagation in a metal (aluminum) cladding waveguide. The TRC technique with a single convolution integral is found to provide higher accuracy, when compared with the recursive convolution counterpart. The methodology is also extended to the unconditionally stable FDTD based on the locally one-dimensional scheme for efficient frequency-dependent calculations.

We propose a new method, called Subclass-oriented Dimensionality Reduction with Pairwise Constraints (SODRPaC), for dimensionality reduction. In a high dimensional space, it is common that a group of data points with one class may scatter in several different groups. Current linear semi-supervised dimensionality reduction methods would fail to achieve fair performances, as they assume two data points linked by a must-link constraint are close each other, while they are likely to be located in different groups. Inspired by the above observation, we classify the must-link constraint into two categories, which are the inter-subclass must-link constraint and the intra-subclass must-link constraint, respectively. We carefully generate cannot-link constraints by using must-link constraints, and then propose a new discriminant criterion by employing the cannot-link constraints and the compactness of shared nearest neighbors. The manifold regularization is also incorporated in our dimensionality reduction framework. Extensive experiments on both synthetic and practical data sets illustrate the effectiveness of our method.

Ultra wideband (UWB) radar is considered a promising technology to complement existing camera-based surveillance systems because, unlike cameras, it provides excellent range resolution. Many of the UWB radar imaging algorithms are based on large-scale antenna arrays that are not necessarily practical because of their complexity and high cost. To resolve this issue, we previously developed a two-dimensional radar imaging algorithm that estimates unknown target shapes and motion using only three antennas. In this paper, we extend this method to obtain three-dimensional images by estimating three-dimensional motions from the outputs of five antennas. Numerical simulations confirm that the proposed method can estimate accurately the target shape under various conditions.

In the present paper, we address the problem of extrapolating group proximities from member relations, which we refer to as the group proximity problem. We assume that a relational dataset consists of several groups and that pairwise relations of all members can be measured. Under these assumptions, the goal is to estimate group proximities from pairwise relations. In order to solve the group proximity problem, we present a method based on embedding and distribution mapping, in which all relational data, which consist of pairwise dissimilarities or dissimilarities between members, are transformed into vectorial data by embedding methods. After this process, the distributions of the groups are obtained. Group proximities are estimated as distances between distributions by distribution mapping methods, which generate a map of distributions. As an example, we apply the proposed method to document and bacterial flora datasets. Finally, we confirm the feasibility of using the proposed method to solve the group proximity problem.

In order to implement multidimensional scaling (MDS) efficiently, we propose a new method named “global mapping analysis” (GMA), which applies stochastic approximation to minimizing MDS criteria. GMA can solve MDS more efficiently in both the linear case (classical MDS) and non-linear one (e.g., ALSCAL) if only the MDS criteria are polynomial. GMA separates the polynomial criteria into the local factors and the global ones. Because the global factors need to be calculated only once in each iteration, GMA is of linear order in the number of objects. Numerical experiments on artificial data verify the efficiency of GMA. It is also shown that GMA can find out various interesting structures from massive document collections.

An iterative inter-track interference (ITI) cancelling scheme is described for multi-track signal detection in nonbinary (NB)-LDPC-coded two-dimensional magnetic recording. The multi-track iterative ITI canceller that we propose consists of multi-track soft interference cancellers (SICs), two-dimensional partial response (TDPR) filters, noise-predictive max-log-MAP detectors, and an NB-LDPC decoder. TDPR filters using an ITI-suppressing tap-weight vector mitigate ITI in the first iteration. Multi-track SICs and TDPR filters adjusted to the residual two-dimensional ISI signals efficiently detect multi-track signals in the latter iterations. The simulation results demonstrated that our proposed iterative multi-track ITI canceller achieves frame error rates close to those obtained in a non-ITI case in media-noise-dominant environments when the both-side off-track ratio is up to 50%.

Telecommunications networks have become an important social infrastructure, and their robustness is considered to be a matter of social significance. Conventional network planning methods are generally based on the maximum volume of ordinary traffic and only assume explicitly specified failure scenarios. Therefore, present networks have marginal survivability against multiple failures induced by an extraordinarily high volume of traffic generated during times of natural disasters or popular social events. This paper proposes a telecommunications network planning method based on probabilistic risk assessment. In this method, risk criterion reflecting the degree of risk due to extraordinarily large traffic loads is predefined and estimated using probabilistic risk assessment. The probabilistic risk assessment can efficiently calculate the small but non-negligible probability that a series of multiple failures will occur in the considered network. Detailed procedures for the proposed planning method are explained using a district mobile network in terms of the extraordinarily large traffic volume resulting from earthquakes. As an application example of the proposed method, capacity dimensioning for the local session servers within the district mobile network is executed to reduce the risk criterion most effectively. Moreover, the optimum traffic-rerouting scheme that minimizes the estimated risk criterion is ascertained simultaneously. From the application example, the proposed planning method is verified to realize a telecommunications network with sufficient robustness against the extraordinarily high volume of traffic caused by the earthquakes.

WSNs (Wireless Sensor Networks) are becoming more widely used in various fields, and localization is a crucial and essential issue for sensor network applications. In this letter, we propose a low-complexity localization mechanism for WSNs that operate in 3D (three-dimensional) space. The basic idea is to use aerial vehicles that are deliberately equipped with anchor nodes. These anchors periodically broadcast beacon signals containing their current locations, and unknown nodes receive these signals as soon as the anchors enter their communication range. We estimate the locations of the unknown nodes based on the proposed scheme that transforms the 3D problem into 2D computations to reduce the complexity of 3D localization. Simulated results show that our approach is an effective scheme for 3D self-positioning in WSNs.

This paper presents a novel method for reconstructing 3D geometry of camera motion and human-face model from a video sequence. The approach combines the concepts of Powell's line minimization with gradient descent. We adapted the line minimization with bracketing used in Powell's minimization to our method. However, instead of bracketing and searching deep down a direction for the minimum point along that direction as done in their line minimization, we achieve minimization by bracketing and searching for the direction in the bracket which provides a lower energy than the previous iteration. Thus, we do not need a large memory as required by Powell's algorithm. The approach to moving in a better direction is similar to classical gradient descent except that the derivative calculation and a good starting point are not needed. The system's constraints are also used to control the bracketing direction. The reconstructed solution is further improved using the Levenberg Marquardt algorithm. No average face model or known-coordinate markers are needed. Feature points defining the human face are tracked using the active appearance model. Occluded points, even in the case of self occlusion, do not pose a problem. The reconstructed space is normalized where the origin can be arbitrarily placed. To use the obtained reconstruction, one can rescale the computed volume to a known scale and transform the coordinate system to any other desired coordinates. This is relatively easy since the 3D geometry of the facial points and the camera parameters of all frames are explicitly computed. Robustness to noise and lens distortion, and 3D accuracy are also demonstrated. All experiments were conducted with an off-the-shelf digital camera carried by a person walking without using any dolly to demonstrate the robustness and practicality of the method. Our method does not require a large memory or the use of any particular, expensive equipment.

In this paper, we propose a two dimensional (2D) non-separable adaptive directional lifting (ADL) structure for discrete wavelet transform (DWT) and its image coding application. Although a 2D non-separable lifting structure of 9/7 DWT has been proposed by interchanging some lifting, we generalize a polyphase representation of 2D non-separable lifting structure of DWT. Furthermore, by introducing the adaptive directional filteringingto the generalized structure, the 2D non-separable ADL structure is realized and applied into image coding. Our proposed method is simpler than the 1D ADL, and can select the different transforming direction with 1D ADL. Through the simulations, the proposed method is shown to be efficient for the lossy and lossless image coding performance.

Histogram-based image features such as HoG, SIFT and histogram of visual words are generally represented as high-dimensional, non-negative vectors. We propose a supervised method of reducing the dimensionality of histogram-based features by using non-negative matrix factorization (NMF). We define a cost function for supervised NMF that consists of two terms. The first term is the generalized divergence term between an input matrix and a product of factorized matrices. The second term is the penalty term that reflects prior knowledge on a training set by assigning predefined constants to cannot-links and must-links in pairs of training data. A multiplicative update rule for minimizing the newly-defined cost function is also proposed. We tested our method on a task of scene classification using histograms of visual words. The experimental results revealed that each of the low-dimensional basis vectors obtained from the proposed method only appeared in a single specific category in most cases. This interesting characteristic not only makes it easy to interpret the meaning of each basis but also improves the power of classification.

We propose a novel method of supervised feature projection called class-distance-based discriminant analysis (CDDA), which is suitable for automatic age estimation (AAE) from facial images. Most methods of supervised feature projection, e.g., Fisher discriminant analysis (FDA) and local Fisher discriminant analysis (LFDA), focus on determining whether two samples belong to the same class (i.e., the same age in AAE) or not. Even if an estimated age is not consistent with the correct age in AAE systems, i.e., the AAE system induces error, smaller errors are better. To treat such characteristics in AAE, CDDA determines between-class separability according to the class distance (i.e., difference in ages); two samples with similar ages are imposed to be close and those with spaced ages are imposed to be far apart. Furthermore, we propose an extension of CDDA called local CDDA (LCDDA), which aims at handling multimodality in samples. Experimental results revealed that CDDA and LCDDA could extract more discriminative features than FDA and LFDA.

This paper proposes a low-complexity signal detection algorithm for spatially correlated multiple-input multiple-output (MIMO) channels. The proposed algorithm sets a minimum mean-square error (MMSE) detection result to the starting point, and searches for signal candidates in multi-dimensions of the noise enhancement from which the MMSE detection suffers. The multi-dimensional search is needed because the number of dominant directions of the noise enhancement is likely to be more than one over the correlated MIMO channels. To reduce the computational complexity of the multi-dimensional search, the proposed algorithm limits the number of signal candidates to O(NT) where NT is the number of transmit antennas and O( ) is big O notation. Specifically, the signal candidates, which are unquantized, are obtained as the solution of a minimization problem under a constraint that a stream of the candidates should be equal to a constellation point. Finally, the detected signal is selected from hard decisions of both the MMSE detection result and unquantized signal candidates on the basis of the log likelihood function. For reducing the complexity of this process, the proposed algorithm decreases the number of calculations of the log likelihood functions for the quantized signal candidates. Computer simulations under a correlated MIMO channel condition demonstrate that the proposed scheme provides an excellent trade-off between BER performance and complexity, and that it is superior to conventional one-dimensional search algorithms in BER performance while requiring less complexity than the conventional algorithms.

Locally linear embedding (LLE) is a well-known method for nonlinear dimensionality reduction. The mathematical proof and experimental results presented in this paper show that the neighborhood sizes in LLE must be smaller than the dimensions of input data spaces, otherwise LLE would degenerate from a nonlinear method for dimensionality reduction into a linear method for dimensionality reduction. Furthermore, when the neighborhood sizes are larger than the dimensions of input data spaces, the solutions to LLE are not unique. In these cases, the addition of some regularization method is often proposed. The experimental results presented in this paper show that the regularization method is not robust. Too large or too small regularization parameters cannot unwrap S-curve. Although a moderate regularization parameters can unwrap S-curve, the relative distance in the input data will be distorted in unwrapping. Therefore, in order to make LLE play fully its advantage in nonlinear dimensionality reduction and avoid multiple solutions happening, the best way is to make sure that the neighborhood sizes are smaller than the dimensions of input data spaces.

Based on the reiterative maximum signal minus interference level (MSMIL) criterion and adaptive beamforming, a novel interference suppression algorithm is proposed for shared-spectrum multistatic radar that must contend with clutter. In this algorithm, two-dimensional adaptive beamformers are designed for azimuths and range cells. Numerical results show advantages of the proposed method.

We propose the collective electron tunneling model in the electron injection process between the Nano Dots (NDs) and the two-dimensional electron gas (2DEG). We report the collective motion of electrons between the 2DEG and the NDs based on the measurement of the Si-ND floating gate structure in the previous studies. However, the origin of this collective motion has not been revealed yet. We evaluate the proposed tunneling model by the model calculation. We reveal that our proposed model reproduces the collective motion of electrons. The insight obtained by our model shows new viewpoints for designing future nano-electronic devices.

This paper presents an improved Gini-Index algorithm to correct feature-selection bias in text classification. Gini-Index has been used as a split measure for choosing the most appropriate splitting attribute in decision tree. Recently, an improved Gini-Index algorithm for feature selection, designed for text categorization and based on Gini-Index theory, was introduced, and it has proved to be better than the other methods. However, we found that the Gini-Index still shows a feature selection bias in text classification, specifically for unbalanced datasets having a huge number of features. The feature selection bias of the Gini-Index in feature selection is shown in three ways: 1) the Gini values of low-frequency features are low (on purity measure) overall, irrespective of the distribution of features among classes, 2) for high-frequency features, the Gini values are always relatively high and 3) for specific features belonging to large classes, the Gini values are relatively lower than those belonging to small classes. Therefore, to correct that bias and improve feature selection in text classification using Gini-Index, we propose an improved Gini-Index (I-GI) algorithm with three reformulated Gini-Index expressions. In the present study, we used global dimensionality reduction (DR) and local DR to measure the goodness of features in feature selections. In experimental results for the I-GI algorithm, we obtained unbiased feature values and eliminated many irrelevant general features while retaining many specific features. Furthermore, we could improve the overall classification performances when we used the local DR method. The total averages of the classification performance were increased by 19.4 %, 15.9 %, 3.3 %, 2.8 % and 2.9 % (kNN) in Micro-F1, 14 %, 9.8 %, 9.2 %, 3.5 % and 4.3 % (SVM) in Micro-F1, 20 %, 16.9 %, 2.8 %, 3.6 % and 3.1 % (kNN) in Macro-F1, 16.3 %, 14 %, 7.1 %, 4.4 %, 6.3 % (SVM) in Macro-F1, compared with tf*idf, χ2, Information Gain, Odds Ratio and the existing Gini-Index methods according to each classifier.

Linear discriminant analysis (LDA) is one of the well-known schemes for feature extraction and dimensionality reduction of labeled data. Recently, two-dimensional LDA (2DLDA) for matrices such as images has been reformulated into symmetric 2DLDA (S2DLDA), which is solved by an iterative algorithm. In this paper, we propose a non-iterative S2DLDA and experimentally show that the proposed method achieves comparable classification accuracy with the conventional S2DLDA, while the proposed method is computationally more efficient than the conventional S2DLDA.

The Network-on-Chip (NoC) is limited by the reliability constraint, which impels us to exploit the fault-tolerant routing. Generally, there are two main design objectives: tolerating more faults and achieving high network performance. To this end, we propose a new multiple-round dimension-order routing (NMR-DOR). Unlike existing solutions, besides the intermediate nodes inter virtual channels (VCs), some turn-legally intermediate nodes inside each VC are also utilized. Hence, more faults are tolerated by those new introduced intermediate nodes without adding extra VCs. Furthermore, unlike the previous solutions where some VCs are prioritized, the NMR-DOR provides a more flexible manner to evenly distribute packets among different VCs. With extensive simulations, we prove that the NMR-DOR maximally saves more than 90% unreachable node pairs blocked by faults in previous solutions, and significantly reduces the packet latency compared with existing solutions.

In the paper, a technique of the numerical inversion of multidimensional Laplace transforms (nD NILT), based on a complex Fourier series approximation is elaborated in light of a possible ralative error achievable. The detailed error analysis shows a relationship between the numerical integration of a multifold Bromwich integral and a complex Fourier series approximation, and leads to a novel formula relating the limiting relative error to the nD NILT technique parameters.