In virtue of great progress in computer graphics technologies, CG movies have been getting popular. However, cinematography techniques, which contribute to improving the contents' comprehensibility, need to be learned from professional experiences, and not easily acquired by non-professional people. This paper focuses on film cutting as one of the most important cinematography techniques in conversational scenes, and presents a system that automatically generates shot transitions to improve comprehensibility of CG contents. First, we propose a cognitive model of User Involvement serving as constraints on selecting shot transitions. Then, to examine the validity of the model, we analyze shot transitions in TV programs, and based on the analysis, we implement a CG contents creation system. Results of our preliminary evaluation experiment show the effectiveness of the proposed method, specifically in enhancing contents' comprehensibility.

This paper presents an analysis of the Slotted DS-CDMA system with modified node components in order to construct a load control structure in which the service rates of each node can be dynamically adapted without using feedback information. In contrast to the traditional Slotted DS-CDMA which is widely represented with single queue, prior emphasis of the approach is laid on the usage of an additional queue which is applied to manage the collided packet traffic while its queue size is also used as a load control parameter. Semi-Markov process is applied to describe the statistic behavior of the system in steady state. Trade-offs between two major performance parameters, i.e., delay and throughput, are presented and compared with those of the traditional system. Results obtained from the simulation and numerical analysis using queuing concept are compared. With these results, an advantage performance for group packets is shown, and we finally extend the concept based on the obtained results to describe a simple algorithm using one way control message as the tool to alleviate the stability problem.

A method is proposed to compute moments of distributed coupled RLC interconnects. Both uniform line models and non-uniform line models will be developed. Considering both self inductances and mutual inductances in multi-conductors, recursive moment computations formulae of lumped coupled RLC interconnects are extended to those of distributed coupled RLC interconnects. By using the moment computation technique in conjunction with the projection-based order reduction method, the inductive crosstalk noise waveform can be accurately and efficiently estimated. Fundamental developments of the proposed approach will be described. Simulation results demonstrate the improved accuracy of the proposed method over the traditional lumped methods.

This paper presents speech-driven Web retrieval models which accept spoken search topics (queries) in the NTCIR-3 Web retrieval task. The major focus of this paper is on improving speech recognition accuracy of spoken queries and then improving retrieval accuracy in speech-driven Web retrieval. We experimentally evaluated the techniques of combining outputs of multiple LVCSR models in recognition of spoken queries. As model combination techniques, we compared the SVM learning technique with conventional voting schemes such as ROVER. In addition, for investigating the effects on the retrieval performance in vocabulary size of the language model, we prepared two kinds of language models: the one's vocabulary size was 20,000, the other's one was 60,000. Then, we evaluated the differences in the recognition rates of the spoken queries and the retrieval performance. We showed that the techniques of multiple LVCSR model combination could achieve improvement both in speech recognition and retrieval accuracies in speech-driven text retrieval. Comparing with the retrieval accuracies when an LM with a 20,000/60,000 vocabulary size is used in an LVCSR system, we found that the larger the vocabulary size is, the better the retrieval accuracy is.

This paper describes an accurate unsupervised speaker adaptation method for lecture style spontaneous speech recognition using multiple LVCSR systems. In an unsupervised speaker adaptation framework, the improvement of recognition performance by adapting acoustic models remarkably depends on the accuracy of labels such as phonemes and syllables. Therefore, extraction of the adaptation data guided by confidence measure is effective for unsupervised adaptation. In this paper, we looked for the high confidence portions based on the agreement between two LVCSR systems, adapted acoustic models using the portions attached with high accurate labels, and then improved the recognition accuracy. We applied our method to the Corpus of Spontaneous Japanese (CSJ) and the method improved the recognition rate by about 2.1% in comparison with a traditional method.

We describe elements of a fast integral equation solver for large periodic and partly periodic finite array systems. A key element of the algorithm is utilization (in a rigorous way) of a block-Toeplitz structure of the impedance matrix in conjunction with either conventional Method of Moments (MoM), Fast Multipole Method (FMM), or Fast Fourier Transform (FFT)-based Adaptive Integral Method (AIM) compression techniques. We refer to the resulting algorithms as the (block-)Toeplitz-MoM, (block-)Toeplitz-AIM, or (block-)Toeplitz-FMM algorithms. While the computational complexity of the Toeplitz-AIM and Toeplitz-FMM algorithms is comparable to that of their non-Toeplitz counterparts, they offer a very significant (about two orders of magnitude for problems of the order of five million unknowns) storage reduction. In particular, our comparisons demonstrate, that the Toeplitz-AIM algorithm offers significant advantages in problems of practical interest involving arrays with complex antenna elements. This result follows from the more favorable scaling of the Toeplitz-AIM algorithm for arrays characterized by large number of unknowns in a single array element and applicability of the AIM algorithm to problems requiring strongly sub-wavelength resolution.

This paper presents a new reconfigurable hardware paradigm, called evolvable hardware (EHW), and its application to the biomedical engineering problem of an artificial hand controller. Evolvable hardware is based on the idea of combining a reconfigurable hardware device with an artificial intelligence robust search technique called genetic algorithms (GAs) to execute reconfiguration autonomously. The first version of the EHW chip was designed in 1998, and this paper describes the latest improvements to the EHW chip, as well as outlining its architecture and the hardware implementation of the GA operations. Execution speed for genetic operations is shown to be about 38.7 times faster with the hardware implementation than with software program running on an AMD Athlon processor (1.2GHz). As an application of the EHW chip, this paper introduces a controller for a multi-functional prosthetic-hand, and presents experimental data in which a practical myoelectric pattern classification rate of 97.8% was achieved through the application of the EHW chip.

In a previous study about a combinatorial optimization problem solver using neural networks, since the Hopfield method, convergence to the optimum solution sooner and with more certainty is regarded as important. Namely, only static states are considered as the information. However, from a biological point of view, dynamical systems have attracted attention recently. Therefore, we propose a "dynamical" combinatorial optimization problem solver using hysteresis neural networks. In this paper, the proposed system is evaluated by the N-Queen problem.

A TCAD implementation of a quantum-mechanical mobility model in the commercial device simulator DESSIS_ISE is presented. The model makes use of an integrated 1D Schrodinger-Poisson solver. Effective mobilities µeff and transfer characteristics are calculated for DGSOI MOSFETs with a wide range of silicon film thickness tSi and buried-oxide thickness tbox. It is shown that the volume-inversion related enhancement of µeff for tSi 10 nm is bound to symmetrical DGSOIs, whereas SIMOX based devices with thick buried oxides limit µeff to the bulk value. The still immature technology makes a conclusive comparison with experimental data impossible.

This paper presents novel techniques for blind separation and blind extraction of instantaneously mixed binary sources, which are suitable for the case with less sensors than sources. First, a solvability analysis is presented for a general case. Necessary and sufficient conditions for recoverability of all or some part of sources are derived. A new deterministic blind separation algorithm is then proposed to estimate the mixing matrix and separate all sources efficiently in the noise-free or low noise level case. Next, using the Maximum Likelihood (ML) approach for robust estimation of centers of clusters, we have extended the algorithm for high additive noise case. Moreover, a new sequential blind extraction algorithm has been developed, which enables us not only to extract the potentially separable sources but also estimate their number. The sources can be extracted in a specific order according to their dominance (strength) in the mixtures. At last, simulation results are presented to illustrate the validity and high performance of the algorithms.

In this paper, a voltage-controlled linear variable resistor (VCLVR) using a floating-gate MOSFET (FG-MOSFET) is proposed. First, the grounded VCLVR realization is discussed. The proposed circuit consists of only an ordinary MOSFET and an FG-MOSFET. The advantages of the proposed VCLVR are low-power and wide-input range and also the power consumption of the proposed VCLVR is the same as an ordinary passive resistor. The performance of the proposed circuits are confirmed by HSPICE simulations with a standard 0.6 µm CMOS process parameters. Simulations of the proposed VCLVR demonstrate a resistance value of 40 kΩ to 338 kΩ and an input range of 4.34 V within THD of less than 1.1%. Next, we proposed a new floating node linear variable resistor using the proposed VCLVR. The performance of the circuit is also evaluated through HSPICE.

In this paper we describe collaborative constraint functional logic programming and the system called Open CFLP that supports this programming paradigm. The system solves equations by collaboration of various equational constraint solvers. The solvers include higher-order lazy narrowing calculi that serve as the interpreter of higher-order functional logic programming, and specialized solvers for solving equations over specific domains, such as a polynomial solver and a differential equation solver. The constraint solvers are distributed in an open environment such as the Internet. They act as providers of constraint solving services. The collaboration between solvers is programmed in a coordination language embedded in a host language. In Open CFLP the user can solve equations in a higher-order functional logic programming style and yet exploit solving resources in the Internet without giving low-level programs of distributions of resources or specifying details of solvers deployed in the Internet.

We have proposed an intracell uplink of a spread-spectrum code-division multiple-access (SS-CDMA) flexible wireless network based on approximately synchronized (AS) CDMA. Since the AS-CDMA has no co-channel interference, complicated transmission power control (TPC) is not required. A modem of the AS-CDMA has been designed and implemented for the Japanese 2.4 GHz industrial, scientific and medical (ISM) band. Using the implemented modem, the degradation of Eb/N0 from the theoretical limit is 1.0 dB at a bit error rate (BER) of 10-3. Under 2-user environment, the degradation of carrier-to-noise ratio (CNR) is 0.5 dB at a BER of 10-3 when the desired-to-undesired signal ratio (DUR) is -20.3 dB. We have evaluated BER performances in cases of varying carrier frequency offset and median DUR with computer simulation. Under 8-user environment, at the carrier frequency offset of 0.3 ppm, the BER with the DUR of -16 dB is found to be 10-3. Using the AS-CDMA with a 4-step open-loop TPC technique, the design of intracell uplink is available.

Generalized Pulse Spectrum Technique (GPST) is a method to solve the inverse problems of wave-propagation and diffusion-dominated phenomena, and therefore has been popularly applied in image reconstruction of time-resolved diffuse optical tomography. With a standard GPST for simultaneous reconstruction of absorption and scattering coefficients, the products of the gradients of the Green's function and the photon-density flux, based on the photon-diffusion equation, are required to calculate the diffusion-related Jacobian matrix. The adversities are of two-folds: time-consuming and singular in the field near the source. The latter causes a severe insensitivity of the algorithm to the scattering changes deep inside tissue. To cope with the above difficulties, we propose in this paper a modified GPST algorithm that only involves the Green's function and the photon-density flux themselves in the scattering-related matrix. Our simulated and experimental reconstructions show that the modified algorithm can significantly improve the quality of scattering image and accelerate the reconstruction process, without an evident degradation in absorption image.

This paper presents an efficient graph-based evolutionary optimization technique called Evolutionary Graph Generation (EGG), and its application to the design of fast constant-coefficient multipliers using parallel counter-tree architecture. An important feature of EGG is its capability to handle the general graph structures directly in evolution process instead of encoding the graph structures into indirect representations, such as bit strings and trees. This paper also addresses the major problem of EGG regarding the significant computation time required for verifying the function of generated circuits. To solve this problem, a new functional verification technique for arithmetic circuits is proposed. It is demonstrated that the EGG system can create efficient multiplier structures which are comparable or superior to the known conventional designs.

Sticking is one of dominant characteristics of reliability in relays for medium current loads from several amperes to several dozen amperes, which are used for relays for automobiles, industrial control units or power supplies of household electrical appliances. Correlations between the release failures due to sticking and contact characteristics such as arc discharges, material parameters and design factors in relays have never been always made clear. This puts difficulty in the way of reasonable development of contact materials and rational design of relays. So, dependence of electrical load conditions on sticking characteristics are investigated, using the Ag-CdO contacts which have had high practical use to relays for medium current loads. Furthermore, relationship among the sticking characteristics, arc discharge characteristics and contact surface properties after operations are studied. Mechanism of sticking is considered on the basis of those data. The results are as follows: (1) Sticking phenomenon occurs intermittently from initial operations and lasts to the end. (2) The µ + 2 σ value (the sum of the mean value and the integral multiple of the standard deviation of sticking force) increases in proportion to the circuit current. On the other hand, it has the maximum value at a circuit voltage, slightly less than the minimum arc voltage. (3) Factors causing the sticking are considered to be divided into direct factors and its root factors. It is considered that a dominant direct factor is welding, and that its root factor is bridge or welding by Joule's heat. On the other hand, the sticking force becomes rather lower as the circuit voltage increases, in the circuit voltage range where regular arc discharge occurs.

In order to examine the validity of Mott-Schottky model at organic/metal interfaces, the position of the vacuum level of N,N'-bis(3-methylphenyl)-N,N'-diphenyl -[1,1'-biphenyl]-4,4'-diamine (TPD) film formed on various metal substrates (Au, Cu, Ag, Mg and Ca) was measured as a function of the film-thickness by Kelvin probe method in ultrahigh vacuum (UHV). TPD is a typical hole-injecting material for organic electroluminescent devices. At all the interfaces, sharp shifts of the vacuum level were observed within 1 nm thickness. Further deposition of TPD up to 100 nm did not change the position of the vacuum level indicating no band bending at these interfaces. These findings clearly demonstrate the Fermi level alignment between metal and bulk TPD solid is not established within typical thickness of real devices.

Design patterns which Erich Gamma advocates is expected as an effective approach for the reuse of designs. So, design patterns are predicted to be used frequently in object-oriented software development. In such circumstance, tools to support applying design patterns to the design diagrams of the system under development are thought to be useful. This research develops Object-Oriented Pattern Applying Support tool OOPAS. It consists of a library of Gamma design patterns with very familiar examples and adrem explanation, and of a function to generate the correctly modified design diagrams of the application system when a design pattern was applied to evolve that system. Actually, these functions are installed in the structured object modeling environment SOME, which is an object-oriented design diagram editor made previously in our laboratory. This design diagram evolving function is formalized as a Join operation of the recursive graph. As a result of the evaluation experiment, the join operation can be applied to the almost of the twenty three Gamma design patterns excluding the six patterns such as Iterator and Command, which are stated at too abstract level to be represented by the design diagrams.

We propose a rough classification system using Hierarchical Learning Vector Quantization (HLVQ) for large scale classification problems which involve many categories. HLVQ of proposed system divides categories hierarchically in the feature space, makes a tree and multiplies the nodes down the hierarchy. The feature space is divided by a few codebook vectors in each layer. The adjacent feature spaces overlap at the borders. HLVQ classification is both speedy and accurate due to the hierarchical architecture and the overlapping technique. In a classification experiment using ETL9B, the largest database of handwritten characters in Japan, (it contains a total of 607,200 samples from 3036 categories) the speed and accuracy of classification by HLVQ was found to be higher than that by Self-Organizing feature Map (SOM) and Learning Vector Quantization methods. We demonstrate that the classification rate of the proposed system which uses multi-codebook vectors for each category under HLVQ can achieve higher speed and accuracy than that of systems which use average vectors.

Excitonic optical properties of GaN homoepitaxial layers have been studied by means of magneto-luminescence and time-resolved luminescence spectroscopy. The luminescence lines due to the radiative recombination of excitons bound to neutral donors and acceptors have been measured under magnetic field up to 8 T, which was aligned perpendicular and parallel to the hexagonal c-axis. Under the magnetic field aligned perpendicular to the hexagonal c-axis, both the donor- and acceptor-bound-exciton lines clearly split into two components, which originated from the Zeeman splitting. The effective g-factors for both the donor- and acceptor-bound excitons were estimated to be 2.02 and 2.47, respectively. Under the magnetic field aligned parallel to the hexagonal c-axis, slight broadening of the bound-exciton lines was observed and the Zeeman splitting was too small to be detected. On the other hand, the diamagnetic shift for both the donor- and acceptor-bound-exciton luminescence lines was observed under the magnetic field aligned both perpendicular and parallel to the hexagonal c-axis. It was found that the diamagnetic shift of the donor-bound exciton was smaller than that of the acceptor-bound exciton. Furthermore, recombination dynamics of excitonic transitions was measured under high-density excitation. An excitation-density-dependent transition of the dominant radiative recombination process from donor-bound excitons to biexcitons was clearly observed in the temporal behavior. In addition, double-exponential decay of biexciton luminescence was observed, which is one of the characteristics of biexciton luminescence at high excitation densities.