Appropriate language modeling is one of the major issues for automatic transcription of spontaneous speech. We propose an adaptation method for statistical language models based on both topic and speaker characteristics. This approach is applied for automatic transcription of meetings and panel discussions, in which multiple participants speak on a given topic in their own speaking style. A baseline language model is a mixture of two models, which are trained with different corpora covering various topics and speakers, respectively. Then, probabilistic latent semantic analysis (PLSA) is performed on the same respective corpora and the initial ASR result to provide two sets of unigram probabilities conditioned on input speech, with regard to topics and speaker characteristics, respectively. Finally, the baseline model is adapted by scaling N-gram probabilities with these unigram probabilities. For speaker adaptation purpose, we make use of a portion of the Corpus of Spontaneous Japanese (CSJ) in which a large number of speakers gave talks for given topics. Experimental evaluation with real discussions showed that both topic and speaker adaptation reduced test-set perplexity, and in total, an average reduction rate of 8.5% was obtained. Furthermore, improvement on word accuracy was also achieved by the proposed adaptation method.

A miniaturized rectangular resonator made of a novel anisotropic artificial dielectric material is investigated which has advantages of small size and big separation of the higher-modes. To obtain a property of anisotropic permittivity, artificial dielectric material is fabricated by lamination of rectangular metal strips etched on a printed circuit board. Artificial dielectric rectangular resonators are designed to excite TE10δ mode selectively, aligning the rectangular metal patterns along the direction of the mode electric field line. The resonant frequencies and coupling coefficient of artificial dielectric rectangular resonators encapsulated in a metal waveguide are analyzed theoretically, and compared with the experimental result. As a microwave application, a high selectivity TE10δ mode bandpass filter (BPF) using two artificial dielectric rectangular resonators is demonstrated. A two-stage BPF with the center frequency of 1.718 GHz, bandwidth of 78 MHz, and insertion loss of 1.3 dB is successfully realized in a rectangular waveguide.

We have determined the thickness and optical constants (refractive index and extinction coefficient) of each layer in the multi-layer organic light emitting diode (OLED) devices based on phosphorescent platinum octaethyl porphine (PtOEP) using a phase modulated spectroscopic ellipsometer. The thickness of each layer estimated from the ellipsometric measurement is different from the thickness measured with quartz oscillator during the evaporation of organic materials. The deviation of total multi-layer thickness is about 5%, while the deviation in each of N, N'-bis(1-naphtyl)-N, N'-diphenyl-1,1'-biphenyl-4,4'-diamine (α-NPD) and aluminum tris 8-hydroxyquinoline (Alq3) layers is about 20-25%. Additionally the spectra of refractive index and extinction coefficient of Alq3 and α-NPD layers are different from those that are measured using the single layer films. These results are understood by penetration of organic material from the neighboring layers in the multi-layer structure devices.

Our goal is to realize an extra-large stereoscopic display in the open air for use by the general public. We have developed a stereoscopic large display by use of a full-color LED panel. Although the developed display enables viewers to view the stereoscopic images without any special glasses, it is necessary for the viewers to move to stand within the viewing areas. Movements of the viewers are considered to depend on arrangements of viewing areas. The purpose of this paper is to investigate the movements of viewers who watch different designs of stereoscopic LED displays with a parallax barrier, including conventional designs to provide multiple perspective images and designs to eliminate pseudoscopic viewing areas, and evaluate the performance of different viewing areas based on the obtained paths of the viewers. We have developed a real-time measurement system of a viewer's position by use of a camera on the ceiling. We have recorded the viewing movements caused by the shift of viewing areas. It was found that the viewers moved to stand on orthoscopic viewing positions. The movements of viewers who move to find a viewing area have been recorded with different designs of stereoscopic LED displays that provide different viewing areas. We have calculated the lateral moving time of the viewers'movements. It is shown that the elimination of pseudoscopic viewing areas reduces the lateral moving time. Thus, the real-time measurement system of a viewer's position has been utilized for evaluation of performance of the different designs of stereoscopic LED displays.

Bluetooth is a system for providing short-range, small size, low-power and low-cost connectivity operating in the ISM (Industrial Scientific Medicine) band at 2.4 GHz. Bluetooth has been seen as a promising candidate for ad-hoc wireless networking and wireless personal area network (WPAN). In this paper, we first discuss previously proposed polling algorithms in Bluetooth piconet. We then propose an efficient fair scheduling algorithm which improves the throughput efficiency of the system by adaptively assigning the polling interval according to the number of inactive slaves. We also show the simulation results of the proposed algorithm compared with previously proposed algorithms.

This paper introduces a methodology for simulating single-electron-transistor (SET)-based multiple-valued logics (MVLs). First, a physics-based analytical model for SET is described, and then a procedure for extracting parameters from measured characteristics is explained. After that, simulated and experimental results for basic MVL circuits are compared. As an advanced example of SET-based logics, a latched parallel counter, which is one of the most important components in arithmetic circuits, is newly designed and analyzed by a simulation. It is found that a SET-based 7-3 counter can be constructed with less than 1/10 the number of devices needed for a conventional circuit and can operate at a moderate speed with 1/100 the conventional power consumption.

Overall picture quality of a liquid crystal display (LCD) is changed by viewing conditions. An evaluation method of overall picture quality was proposed for LCDs. However, the estimation values calculated by the evaluation method differed from the subjective evaluation values, when viewing angle changed in the vertical direction. In this study, the coefficients of impairment in the evaluation equation were obtained by the multiple regression analysis using subjectively evaluated data. The evaluation method of overall picture quality was improved by weighting factors using these coefficients. As the result, the good estimation accuracy was obtained even if the viewing angle was changed both in the horizontal and vertical directions individually.

An iterative method is proposed to solve integral equations (IEs) of the second kind with Picard-kernel in linear complexity, i.e.O(N). The particular IE considered describes the process of scattering of a plane wave incident on an inhomogeneous slab. The collocation method with triangle basis functions is used to derive a linear system of equations, which is solved for a test problem with the BiCGSTAB method. To reduce the number of iterations, an efficient preconditioning operator is introduced.

Digital watermarks on pictures must have the ability to survive various image processing operations while not causing degradation of picture quality. Random geometric distortion is one of the most difficult kinds of image processing for a watermark to survive, and this problem has become a central issue in watermarking research. Previous methods for dealing with random geometric distortion have been based on searches, special watermark patterns, learning, or additional data such as original pictures. Their use, however, is accompanied by large computational overhead or by operational inconvenience. This paper therefore proposes a method based on embedding watermark patterns in two of the three color planes constituting a color picture so that these two planes have a specific covariance. The detection of the embedded information is based on the covariance between these two planes. Random geometric distortion distorts all the constituent color planes of a picture in the same way and thus does not affect the covariance between any two. The covariance-based detection is therefore immune to the distortion. The paper clarifies that detection error would occur whenever the inherent covariance (the covariance in the original picture) overrides the covariance made by watermarking. The two constituent planes having the minimum inherent covariance are therefore selected and their inherent covariance is reduced by shifting one of them and using a noise-reduction preprocess. Experimental evaluations using StirMark confirmed that 64 bits embedded in 256256-pixel pictures can be correctly detected without using searches, special patterns, learning, or additional data.

We propose a new method of direct mode motion compensation for bi-directionally predicted pictures (B-pictures). The proposed direct mode system is utilized in extended multiple picture prediction, in which blocks in B-pictures are encoded by referring to previous B-pictures in addition to I- or P-pictures as forward reference pictures in a multiple picture prediction framework. The proposed direct mode is suitable for extended multiple picture prediction, since it always uses the immediately previous picture as the forward reference picture. In the simulation, our proposed method is implemented in the H.26L codec. The simulation results show that the extended multiple picture prediction employing the proposed direct mode can reduce the bit rate of B-pictures by up to nearly 13% compared to conventional multiple picture prediction under typical encoding conditions. We also evaluate the performance of the proposed direct mode with the extended multiple picture prediction under several different encoding conditions.

This paper proposes a PCB plane model for generic circuit simulators (SPICE). The proposed model reflects two frequency-dependent losses, namely, skin and dielectric losses. Once power/ground plane pair is divided into arrays of unit-cells, each unit-cell is modeled using a transmission line and the loss model. The loss model is composed of a resistor for DC loss, series RL ladder circuit for skin loss and series RC ladder circuit for dielectric loss. To verify the validity of the proposed model, it is compared with SPICE ac analysis using frequency-dependent resistors. Also, we show that the estimation results using the proposed model have a good correlation with that of VNA measurement for the typical PCB stack-up structure of general desktop PCs. With the proposed model, not only ac analysis but also transient analysis can be easily done for circuits including various non-linear/linear devices since the model consists of passive elements only.

In this letter, adaptive multi-stage parallel interference cancellation (PIC) receiver is considered for multi-rate DS-CDMA system. In each stage of the adaptive multi-stage PIC receiver, multiple access interference (MAI) estimates are obtained by the sub-bit estimates from the previous stage and the adaptive weights for the sub-bit estimates. The adaptive weights are obtained by minimizing the mean squared error between the received signal and its estimate through normalized least mean square (LMS) algorithm. It is shown that the adaptive multi-stage PIC receiver achieves smaller BER than the matched filter receiver, multi-stage PIC receiver, and multi-stage partial PIC receiver for the multi-rate DS-CDMA system in a Rayleigh fading channel.

Novel designs of multiple-valued logic (quaternary) half adder, full adder, and carry-lookahead adder are introduced. The proposed circuits are static and operate in voltage-mode. Moreover, there is no current flow in steady states, and thus, no static power dissipation. Although the comparison in transistor count shows that the proposed quaternary circuits are larger than two respective binary ones, benefits in parallel addition arise from the use of multiple-valued logic. Firstly, the ripple-carry additions are faster because the number of carries is half compared to binary ones and the propagation delay from the input carry through the output carry is relatively small. Secondly, the carry-lookahead scheme exhibits less complexity, which leads to overall reduction in transistor count for addition with large number of bits.

In texture mapping, anisotropic filtering methods, which require more texels, have been proposed for high-quality images. Memory bandwidth, however, is still limited by a bottleneck in the texture-filtering hardware. In this paper, we propose anisotropic texture filtering based on edge function. In generating the weight that plays a key role in filtering texels loaded from memory, the edge function gives accurate contribution of texels to the pixel intensity. The quality of images is superior to other methods. For images of the same quality, our method requires less than half the texels of other methods. In other words, the improvement in performance is more than twice that of other methods.

This paper presents practical modeling procedure of feed patterns, bond wires, and interconnects for microwave bare-chip devices. Dedicated test structures have been designed for the process. Modeling accuracy of BJTs and diodes has been unprecedentedly improved up to 30 GHz with this procedure despite popular SPICE models were used.

This paper describes a sparse realization of passive reduced-order interconnect models via PRIMA to provide the SPICE compatible models. It is demonstrated that, if the SPICE models are directly realized so that the reduced-order equations obtained via PRIMA are stamped into the MNA matrix, the simulations of networks containing the macromodels become computationally inefficient when size of the reduced-order equations is relatively large. This is due to dense coefficient matrices of the reduced-order equations resulting from congruent transformations in PRIMA. To overcome this disadvantage, we propose a sparse realization of the reduced-order models. Since the expression is equivalent to the reduced-order equations, the passivity of the SPICE models generated is also guaranteed. Computational efficiency on SPICE is demonstrated in performing the transient analysis of circuits containing the proposed macromodels.

It is inevitable for driver assist and warning systems to consider the drivers' state of consciousness. Drowsiness is one of the important factors in estimating the drivers' state of consciousness. A Method to extract the driver's initial stage of drowsiness was developed by means of the eyelid's opening relevant to each various characteristic of objects with motion pictures processing in the actual driving environment. The result was that an increase of the long eyelid closure time was the key factor in estimating the initial stage of drivers' drowsiness while driving. And the state of drowsiness could be presumed by checking the frequencies of long eyelid closure time per unit period.

The size of the microscopic electron spot is an important parameter for the white-uniformity of a CRT. It changes as a function of the focus voltage and beam repulsion. This paper explains the mechanism behind this phenomenon. The model is supported by means of measurements.

Making up an electromagnetic wave simulator based on the FEM is tried, which may run on some widely used platforms by use of Java and a single commercial tool. Since the codes and configuration files to be created for this simulator are common, one can construct the simulator running on the platforms at the same time. Using this simulator, the transmission properties of two- and three-dimensional waveguide discontinuities in optical and microwave waveguides are analyzed, the inverse problem in material constant measurement is solved, and the computed results are presented including plots of the electric field distribution.

Edge detection has been an essential step in image processing, and there has been much work undertaken to date. This paper inspects a fuzzy mathematical morphology in order to reach a higher-level of edge-image processing. The proposed scheme uses a fuzzy morphological gradient to detect object boundaries, when the boundaries are roughly defined as a curve or a surface separating homogeneous regions. The automatic edge detection algorithm consists of two major steps. First, a new version of anisotropic diffusion is proposed for edge detection and image restoration. All improvements of the new version use fuzzy mathematical morphology to preserve the edge accuracy and to restore the images to homogeneity. Second, the fuzzy morphological gradient operation detects the step edges between the homogeneous regions as object boundaries. This operation uses geometrical characteristics contained in the structuring element in order to extract the edge features in the set of edgeness, a set consisting of the quality values of the edge pixels. This set is prepared with fuzzy logic for decision and selection of authentic edge pixels. For experimental results, the proposed method has been tested successfully with both synthetic and real pictures.