The conventional shape from focus (SFF) methods have inaccuracies because of piecewise constant approximation of the focused image surface (FIS). We propose a more accurate scheme for SFF based on representation of three-dimensional FIS in terms of neural network weights. The neural networks are trained to learn the shape of the FIS that maximizes the focus measure.

This paper describes the architecture to implement an application in network environments, which adapts to unexpected change in the development phase. In this architecture, an application is expressed as an agent which consists of two layers: base level and meta level. The base level program is an application program and the meta level program is the program that controls the execution of the base level and changes the base level program. Virtual places are also provided in the network. They are used for the release of programs and information which agents retrieve to change their own base level program. An application (or an agent), when a change is required, moves from places to places for the retrieval of programs to adapt to the change. A program search strategy is introduced to adapt to changes by using distributed thesauri of released programs, which realizes an agent's program retrieval method in network environments.

This paper proposes a fast method of calculating high-order backward Linear Prediction (LP) coefficients for wideband Code Excited LP (CELP) coders operating at around 16 kbit/s. The fast calculation is achieved by a recursive calculation for the high-order autocorrelation of the decoded signal. The recursive calculation can be employed thanks to a novel method of converting the autocorrelation of the decoded signal to that of the residual signal. High-order backward LP coefficients are computed from the autocorrelation of the residual signal using the Levinson-Durbin (LD) procedure. The conversion approximately performs inverse-filtering using LP coefficients representing a corresponding envelope spectrum. Due to the recursive calculation, the proposed fast calculation method achieves 30% to 45% reduction in computations to calculate the high-order backward LP coefficients compared to the conventional method. Subjective tests show that a wideband Multi-Pulse based CELP (MP-CELP) coder at 16 kbit/s with the proposed method achieves comparable coding quality to that with the conventional one with 35% reduction in computations needed for calculation of the backward LP coefficients.

Let l be a positive integer, and let G be a graph with nonnegative integer weights on edges. Then a generalized vertex-coloring, called an l-coloring of G, is an assignment of colors to the vertices of G in such a way that any two vertices u and v get different colors if the distance between u and v in G is at most l. In this paper we give an algorithm to find an l-coloring of a given graph G with the minimum number of colors. The algorithm takes polynomial time if G is a partial k-tree and both l and k are bounded integers.

This note investigates the characterizing properties of the level sets of an M-convex function introduced by Murota.

PID control schemes have been widely used in most of process control systems. Most of these processes are often treated as first-order systems with dead times. And also, in many cases, PID parameters are usually tuned based on the process parameters, i. e. , the time constant, the dead time and the process gain. However, since these process parameters can not be obtained exactly, it is well known that it is difficult to find the suitable PID parameters in practice. In this paper, we propose a genetic tuning scheme of PID parameters for first-order systems with large dead times. The authors have already proposed a tuning method of PID parameters using a genetic algorithm (GA), which was based on the relationship between PID control and generalized minimum variance control(GMVC) laws. In practice, for large dead time systems, first-order low pass pre-filters are often used. The proposed method is an extended version of the previously proposed method mentioned above to the system with a pre-filter due to the large dead time, i. e. , a tuning method of both PID parameters and the pre-filter using a GA. The proposed control scheme is numerically evaluated on some simulation examples.

This paper proposes and evaluates a character or symbol code system called EPICS for multilingual information processing. EPICS integrates a variable-length coding system using 16-bit units and a smart virtual machine. EPICS enhances the interchangeability of data. The variable-length coding system provides a huge code space. This huge space can include not only standardized code sets but also non-standardized codes of ancient symbols and user-specific symbols. The smart virtual machine executes inputs as instructions and is dynamically customizable. It allows us to define and modify instructions during runtime and provides us with customization facilities. Customization facilities can be used to specify a sorting order and normalization. Customization also makes it possible for an information producer (sender) to express his intentions or annotations in data and for an information consumer (receiver) to process the data depending on his needs. Moreover, customization enables one to send compressed data and decompression program fragments incrementally and efficiently without predefined decompression algorithms.

There are many public key cryptosystems that require random inputs to encrypt messages and their security is always discussed assuming that random objects are ideally generated. Since cryptosystems run on computers, it is quite natural that these random objects are computationally generated. One theoretical solution is the use of pseudorandom generators in the Yao's sense. Informally saying, the pseudorandom generators are polynomial-time algorithms whose outputs are computationally indistinguishable from the uniform distribution. Since if we use the Yao's generators then it takes much more time to generate pseudorandom objects than to encrypt messages in public key cryptosystems, we relax the conditions of pseudorandom generators to fit public key cryptosystems and give a minimal requirement for pseudorandom generators within public key cryptosystems. As an example, we discuss the security of the ElGamal cryptosystem with some well-known generators (e. g. , the linear congruential generator). We also propose a new pseudorandom number generator, for random inputs to the ElGamal cryptosystem, that satisfies the minimal requirement. The newly proposed generator is based on the linear congruential generator. We show some evidence that the ElGamal cryptosystem with the proposed generator is secure.

Formal description techniques (FDTs) such as VDM, Z, LOTOS, etc are powerful to develop safety-critical systems since they have strict semantics and mathematical reasoning basis. However, they have no methods or guides how to construct specifications unlike specification and design methods such as Object-Oriented Modeling and Technique (OMT), and that makes it difficult for practitioners to compose formal specifications. One of the solutions is to connect formal description techniques with some existing methods. This paper discusses a technique how to integrate FDTs with specification and design methods such as OMT so that we can have new methods to support writing formal specifications. The integration mechanism is based on transformation rules of specification documents produced following methods into the descriptions written in formal description techniques. The transformation rules specify the correspondences on two meta models; of methods and of formal description techniques, and are described as graph rewriting rules. As an example, we pick up OMT as a method and LOTOS as a FDT and define the transformation rule on their meta models.

In order to get a perspective to the future of GaN materials, theoretical and experimental knowledge of the optoelectronic activities of dislocations in hexagonal GaN have been reviewed. Although the dislocations in GaN have been thought to be not quite harmful, a growing number of evidences have been accumulated for the intrinsic noxiousness of the dislocations. There are some inconsistencies between experimental data reported by different groups or at different dates, which can be reconciled by a proposed simple model that takes into account the trapping of free excitons. A transmission electron microscopic study revealed that some type of dislocations exhibit the recombination enhanced dislocation glide effect, suggesting the non-radiative recombination activity of the fresh dislocations. Such intrinsic activities of dislocations in GaN, in both electronical and mechanical respects, will possibly cause great difficulties in optoelectronic devices based on this material when the crystal quality becomes improved.

This work presents a scalable and high performance prediction protocol for optical networks. In the proposed protocol, we develop a mathematical model to maintain the stability of a network system by prediction based on the traffic temporal locality property. All the critical factors, including transceiver tuning time, propagation delay, and processing time for dealing with control packets, are considered in the proposed prediction protocol. Furthermore, our protocol can resolve the bottlenecks attributed to control signaling and electronics processing. The performance evaluation reveals that the proposed scheme can yield the higher bandwidth efficiency and incur a lower packet delay than those of the TDM and conventional reservation schemes. Also, the proposed protocol can flexibly support any scaled network system such as MANs or LANs.

In this paper, we propose a method for detecting conserved domains from a set of amino acid sequences that belong to a protein family. This method detects the domains as follows: first, generate fixed-length subsequences from the sequences; second, construct a weighted graph that connects any two of the subsequences (vertices) having higher similarity than a pre-defined threshold; third, search for the maximum-density subgraph for each connected component of the graph; finally, explore conserved domains in the sequences by combining the results of the previous step. From the performance results obtained by applying the method to several protein families that have complex conserved domains, we found that our method was able to detect those domains even though some domains were weakly conserved.

The long spikes have been often recorded at the multiples of the electron cyclotron frequency in the ionograms of the topside sounders observed in low latitudes. There has not been sufficient explanation for the physical cause for occourrence of the long spike so far. Here, by interpreting this phenomenon as receiving the trapped cyclotron harmonic wave, some analyses for the length of spike are done not only from the viewpoint of the sweeping property of the frequency spectrum of the transmitted pulse but also from that of the mutual positional relation between the propagation path and the orbit of the sounder. The cause of forming a single spike and a graphical calculation method for the long spike are proposed, respectively. Thus, the cause and the fine structure of long spike consisting of superposed spikes are clarified.

The relationship between the change in transistor operation regions and the fault behavior of a mixed-signal circuit having a bridging fault was investigated. We also discussed determination of transistors to be observed for estimating the fault behavior. These results will be useful for modeling faulty behaviors and analyzing and diagnosing faults in mixed-signal circuits.

This paper proposes a method of automatically eliciting knowledge which is used to detect feature interactions in telecommunication services. With conventional methods, the knowledge is provided manually. With the proposed method, the knowledge is automatically elicited as service constraints. In telecommunication systems, when a new service is added, new state transitions are created. In case of new service, the new state should be reached in the state transitions. On the other hand, some states of existing services should not be reached. These constraints can be considered as knowledge for detecting feature interactions. This paper also proposes a scenario for detecting feature interactions using elicited knowledge. This scenario was confirmed as effective.

Gaze detection is to locate the position on a monitor screen where a user is looking. In our work, we implement it with a computer vision system setting a single camera above a monitor and a user moves (rotates and/or translates) her face to gaze at a different position on the monitor. For our case, the user is requested not to move pupils of her eyes when she gazes at a different position on the monitor screen, though we are working on to relax this restriction. To detect the gaze position, we extract facial features (both eyes, nostrils and lip corners) automatically in 2D camera images. From the movement of feature points detected in starting images, we can compute the initial 3D positions of those features by recursive estimation algorithm. Then, when a user moves her head in order to gaze at one position on a monitor, the moved 3D positions of those features can be computed from 3D motion estimation by Iterative Extended Kalman Filter (IEKF) and affine transform. Finally, the gaze position on a monitor is computed from the normal vector of the plane determined by those moved 3D positions of features. Especially, in order to obtain the exact 3D positions of initial feature points, we unify three coordinate systems (face, monitor and camera coordinate system) based on perspective transformation. As experimental results, the 3D position estimation error of initial feature points, which is the RMS error between the estimated initial 3D feature positions and the real positions (measured by 3D position tracker sensor) is about 1.28 cm (0.75 cm in X axis, 0.85 cm in Y axis, 0.6 cm in Z axis) and the 3D motion estimation errors of feature points by Iterative Extended Kalman Filter (IEKF) are about 2.8 degrees and 1.21 cm in rotation and translation, respectively. From that, we can obtain the gaze position on a monitor (17 inches) and the gaze position accuracy between the calculated positions and the real ones is about 2.06 inches of RMS error.

Dislocation properties in InGaN/GaN Quantum Wells and GaN grown on bulk GaN and sapphire substrates by metalorganic chemical vapor deposition (MOCVD) were characterized using cathodoluminescnece (CL), transmission electron microscopy (TEM), atomic force microscopy (AFM) and photoluminescence (PL). It was clearly demonstrated that dislocations act as nonradiative recombination centers in both n-type (undoped and Si-doped) GaN and InGaN layers. Furthermore the very short-minority carrier diffusion length was a key parameter to explain the high light emission efficiency in GaN-based light emitting diodes (LEDs) prepared on sapphire substrates. On the other side band-tail states were detected in the heteroepitaxial InGaN layers only by temperature dependence PL measurement. Additionally InGaN phase separation, which consists of few micron domains, has been produced under growth conditions which favors the spiral growth. These results indicate that the dislocations in the InGaN layers act as triggering centers for the InGaN phase separation which cause both a compositional fluctuation and the formation of few micron phase separated domains. The homoepitaxial InGaN layers showed however quite normal behaviors for all characterizations.

Investigations were carried out on metalorganic-chemical-vapor-deposition (MOCVD)-grown strained AlGaN/ GaN/sapphire structures using X-ray diffratometry. While AlGaN with lower AlN molar fraction (< 0.1) is under the in-plane compressive stress, it is under the in-plane tensile stress with high AlN molar fraction (> 0.1). Though tensile stress caused the cracks in AlGaN layer with high AlN molar fraction, we found that the cracks dramatically reduced when the GaN layer quality was not good. Using this technique, we fabricated a GaInN multi-quantum-well (MQW) surface emitting diodes were fabricated on 15 pairs of AlGaN/GaN distributed Bragg reflector (DBR) structures. The reflectivity of 15 pairs of AlGaN/GaN DBR structure has been shown as 75% at 435 nm. Considerably higher output power (1.5 times) has been observed for DBR based GaInN MQW LED when compared with non-DBR based MQW structures.

We studied Si and Mg doping characteristics in cubic GaN and fabricated a light emitting diode of cubic GaN on a GaAs substrate by metalorganic vapor-phase epitaxy. The diode structure consisted of undoped and Mg-doped GaN stacking layers deposited on Si-doped GaN and AlGaN layers. The electron-beam-induced-current signal and current injection characteristics of this diode structure were measured. There was a peak at the interface between the Mg-doped and undoped GaN in the electron-beam-induced-current signal. This shows successful growth of the p-n junction. Light emitting operation was achieved by currents injected through the conducting GaAs substrate of this diode at room temperature. We observed electroluminescence below the bandgap energy of cubic GaN with a peak at 2.6 eV.

Our recent progress in GaN-based nanostructures for quantum dot (QD) lasers and vertical microcavity surface emitting lasers (VCSELs) is discussed. We have grown InGaN self-assembled QDs on a GaN epitaxial layer, using atmospheric-pressure metalorganic chemical vapor deposition. The average diameter of the QDs was as small as 8.4 nm and strong photoluminescence emission from the QDs was observed at room temperature. Furthermore, we found that InGaN QDs could be formed even after 10 QD layers were stacked, thus increasing the total QD density. Using these growth results, we fabricated a laser structure with InGaN QDs embedded in the active layer. A clear threshold was observed in the dependence of the emission intensity on the excitation energy at room temperature under optical excitation. We succeeded in demonstrating in lasing action in vertical cavity surface emitting lasers at room temperature with a cavity finesse of over 200.