A new algorithm for texture segmentation, called iterative feature extraction (IFE), is proposed to iteratively search and select for an overcomplete wavelet feature vector based on aspect ratio of extrema number (AREN) with a desired window that provides optimal classification accuracy.

We have developed Thai speech synthesis by rule using cepstral parameters. In order to synthesize the pitch contour of Thai tones, we have applied an extension of Fujisaki's model. A mid tone is unique for Thai when compared to Chinese. For the extension of Fujisaki's model to Thai tones, we assumed that the mid tone is neutral and we adopted its phrase component as the phrase components for all tones. According to our study on the pitch contour of five Thai tones using this model, the result shows that the command pattern for the local F0 components needs both positive and negative commands. Listening tests showed that the intelligibility of the Thai tones measured in terms of error rate were 0.0%, 0.7% and 2.7% for analysis/synthesis, Fujisaki's model and the polynomial model, respectively. Therefore, it is shown that the extension of Fujisaki's model is effective for Thai.

Automatic image inspector inspects the quality of printed circuit boards using image-processing technology. In this study, we change an automatic inspection problem into a problem for detecting the signal singularities. Based on the wavelet theory that the wavelet transform can focus on localized signal structures with a zooming procedure, a novel singularity detection and measurement algorithm is proposed. Singularity positions are detected with the local wavelet transform modulus maximum (WTMM) line, and the Lipschitz exponent is estimated at each singularity from the decay of the wavelet transform amplitude along the WTMM line. According to the theoretical analysis and computer simulation results, the proposed algorithm is shown to be successful for solving the automatic inspection problem and calculating the Lipschitz exponents of signals. These Lipschitz exponents successfully characterize singular behavior of signals at singularities.

We study the advantages and limitations of InP/GaAsSb/InP DHBTs for high-speed digital circuit applications. We show that the high-current performance limitation in these devices is electrostatic in nature. Comparison of the location of collector current blocking in various collector designs suggests a smoother, more gradual onset of blocking effects in type-II collectors. A comparison of collector current blocking effects between InP/GaAsSb--based and various designs of InP/GaInAs--based DHBTs provides support for our analysis.

A Time Hopping Pulse Spacing Modulation (TH-PSM) system, which combines the pulse position modulation system with code shift keying, is proposed. The following performances are analyzed; (1) data transmission rate, (2) error rate in a single-user case, (3) error rate in a multi-user case, and (4) spectral efficiency. Consequently, the data transmission rate of the proposed system is higher than that of the conventional Spread Spectrum Pulse Position Modulation (SS-PPM) system. The proposed system can improve the probability of block error by increasing the number of information bits per spreading code. Moreover, the spectral efficiency of the proposed system is higher than that of the conventional system. The proposed system is more attractive than the conventional SS-PPM system for optical communications, power-line communications, and UWB communications.

The DC resistivities of silicon germanium thin films on Si substrates by a non-contact and non-destructive technique using terahertz time domain spectroscopy (THz-TDS) agree with the values obtained by the four-point probe measurement. In the present experiment, the mobility has not been precisely determined owing to the limitation of the frequency range in our equipment (from 0.1 to 1.5 THz). However, when the mobility becomes large enough, this method will be highly useful in evaluating semiconductor thin films, since the method gives the same data as those from Hall measurement without sample processing or electrode contact to sample.

In this paper, we describe the operation of circuits capable of more than 40-Gbit/s that we have developed using InP HEMT technology. For example, we succeeded in obtaining 43-Gbit/s operation for a full-rate 4:1Multiplier (MUX), 50-Gbit/s operation for a Demultiplexer (DEMUX), 50-Gbit/s operation for a D-type flip-flop (D-FF), and a preamplifier with a bandwidth of 40 GHz. In addition, the achievement of 90-Gbit/s operation for a 2:1MUX and a distributed amplifier with over 110-GHz bandwidth indicates that InP HEMT technology is promising for system operations of over 100 Gbit/s. To achieve these results, we also developed several design techniques to improve frequency response above 80 GHz including a symmetric and separated layout of differential elements in the basic SCFL gate and inverted microstrip.

In generating Frequency-Hopping (FH) sequences for Frequency Hopping Spread Spectrum Multiple Access (FH-SSMA) applications, binary maximal-length sequences (m-sequences) over GF(2m) have been preferred because of their characteristics of good Hamming correlation property, long period capability and high speed generation in association with simple hardware implementation based on Feedback Shift Registers (FSR). In practice, however, one difficulty of applying such sequences into the FH-SSMA communication systems with a wide bandwidth such as Military Satellite Communication (MilSatCom) is that the number of hopping frequency slots available may be far from a power of 2 in proportion to the spreading bandwidth. In that case, we can not make good use of the spreading bandwidth. In this paper, we propose a construction of some favorable FH sequences which deals effectively with the above difficulty using FSR and some nonlinear logic by introducing a re-mapping method. We show that the resulting sequences satisfy the (almost) uniform symbol distribution in one period and preserve a reasonably good Hamming correlation property so that they are appropriate for FH-SSMA applications.

This paper describes a method of analyzing musical sound using a self-organizing map. To take compound factors into account, energy spectra whose frequency ranges were based on the psycho-acoustic experiments were used as input data. Results for music compact discs confirmed that our method could effectively display the positioning and relationships among musical sounds on a map.

Whispering-Gallery mode resonator method has been presented for complex permittivity evaluation of low loss dielectric materials in millimeter wave region. As a problem, it has been found that the evaluation error slightly dependens on the frequency for a sample. It comes from approximated analysis which is used in the procedure. In this paper, a mode-matching method is applied to this evaluation technique to have improvement of the measrued results. It is confirmed experimentally that reliability of the presented method is improved for the millimeter wave permittivity measurement.

In this paper a digital frequency domain RFI (Radio Frequency Interference) cancellation scheme for DMT (Discrete Multitone) based VDSL (Very high speed Digital Subscriber Line) systems is presented. The proposed algorithm has been optimized and characterized in terms of complexity and performance. Optimizations were also performed from an implementation point of view by deducing key dependencies among our RFI model coefficients that let us drastically reduce the size of the memories involved. System simulations showed the effectiveness of the canceller: in terms of VDSL performance parameters such as bit rate, the optimized cancellation scheme recovers almost totally the performance degradation due to RFI.

Adaptive arrays have been recognized as an attractive mean for overcoming multipath fading and interference rejection in the field of mobile communications. In, an adaptive array applicable to single-user spread spectrum (SS) systems has been proposed. In this system, any a priori information concerning incoming signal, even the spreading code and synchronization, is not necessary while it achieves equalizing, beamforming and despreading of a received signal, simultaneously. In this paper, we propose a multistage blind adaptive array antenna based on the above-mentioned adaptive algorithm in order to realize blind signal processing that is applicable to multi-user SS systems. Behavior and performance of the proposed multistage system are examined through computer simulations.

This paper presents a frequency domain simultaneous equations method capable of automatically recovering noise reduction effect degraded by secondary path changes. The simultaneous equations method has been studied, first in time domain. Accordingly to the study, in the time domain, the simultaneous equations method requires an additional filter and a system identification circuit used for transforming the solution of the simultaneous equations into the coefficients of noise control filter, which increase the processing cost. To reduce the processing cost, this paper studies on the application of a frequency domain processing technique, the cross spectrum method, to the simultaneous equations method. By directly applying the equation defining the cross spectrum method to the solution, the additional filter becomes unnecessary. In addition, the system identification circuit is replaced with the inverse Fourier transform. Thereby, the processing cost drastically decreases. This paper also presents simulation results to confirm that the proposed method can automatically recover the noise reduction effect degraded by a path change and provides much higher convergence speed than that of the filtered-x NLMS algorithm with the perfectly modeled secondary path filter.

In this paper, a practical adaptive TuCM scheme is proposed, and its adaptive method is described. With some hardware considerations, a suboptimal optimization algorithm which shows that the number of fading regions is variable is put forward. The proposed adaptive TuCM comes within 3 dB of fading channel capacity, exhibits about 3 dB power gain over conventional adaptive TCM, and is easy to realize by hardware. Considering delay and channel estimation error, the BER performance of adaptive TuCM is analyzed and simulated. In the performance analysis, the method of data fitting is applied to obtain the BER expression for TuCM, and a fitting mathematical model is proposed. Results show that adaptive TuCM is very sensitive to delay and channel estimation error. To alleviate these problems, we proposed an improved power adaptation that can make adaptive TuCM practical.

This paper describes a new training method of average voice model for speech synthesis in which arbitrary speaker's voice is generated based on speaker adaptation. When the amount of training data is limited, the distributions of average voice model often have bias depending on speaker and/or gender and this will degrade the quality of synthetic speech. In the proposed method, to reduce the influence of speaker dependence, we incorporate a context clustering technique called shared decision tree context clustering and speaker adaptive training into the training procedure of average voice model. From the results of subjective tests, we show that the average voice model trained using the proposed method generates more natural sounding speech than the conventional average voice model. Moreover, it is shown that voice characteristics and prosodic features of synthetic speech generated from the adapted model using the proposed method are closer to the target speaker than the conventional method.

This paper proposes a method to integrate computer specifications retrieved from multiple Web sites, to extract characteristic-data of each computer based on integrated information, and to present products suitable for a user's request. The specifications written in HTML are converted into normal forms called table structure. The quantitative attributes such as speed, capacity and dimensions are extracted by comparing them with the mean or mode of all sample data, and the qualitative ones such as kind of processor and graphics chip are extracted using knowledge provided manually. The recommended products are dynamically determined from the extracted data by a user's request and relevance feedback. Experimental results show the effectiveness of our method.

This paper presents a new modulation scheme for Very-High Speed Digital Subscriber Lines (VDSL) modem, featuring a Multi-Code Multi-Carrier Code Division Multiple Access (MC2-CDMA) modulation. The system takes advantage from both the CDMA modulation and the Multi-Carrier transmission, and furthermore the channel throughput is increased adopting a multi-code approach. Starting from an overview of this novel scheme, encompassing the transmitter, channel and receiver description, a brief review of the equalization techniques is also considered and a proper bit-loading algorithm is derived to find out the achievable overall channel rate. The aim of this paper, besides introducing this novel scheme, is to demonstrate its suitability for a VDSL environment, where the achievable channel rate represents a real challenge. By means of a further optimisation, a general improvement of the system performance with respect to the standardized Discrete Multi Tone (DMT) modulation is also demonstrated.

The millimeter wave bandpass filter using the Whispering-Gallery mode (WG mode) dielectric disk resonators is presented in this paper. A 4 stage maximally flat bandpass filter is constructed with the PTFE disk resonators. For the filter design, the coupling coefficients of this mode in the coupled disk resonators are calculated by an approximated separation of variables method. Furthermore, the external Q values of the disk resonator excited by a dielectric waveguide are investigated experimentally. Designed center frequency is 60 GHz and 3 dB band width is 150 MHz. Furthermore, as an attempt to improve the spurious characteristics, another filter structure which consists of some kinds of dielectric disk is tested. As a result, some spurious responses are reduced considerably.

We demonstrate polarization-independent and highly-efficient optical fiber wavelength converters in a 10 Gb/s NRZ transmission system. They are based on synchronous phase or frequency modulations of the two orthogonally polarized pump lights, and can suppress not only the stimulated Brillouin scattering (SBS) but also the spread of the converted spectrum without modulating the signal light.

A new technique for optical generation of high-purity millimeter-wave (mm-wave) signals--namely, by synthesizing the outputs from cascadingly phase-locked multiple semiconductor lasers--was developed. Firstly, a high-spectral-purity mm-wave signal was optically generated by heterodyning the outputs from two phase-locked external-cavity semiconductor lasers. The beat signal was detected by a p-i-n photodiode whose output was directly coupled to a coax-waveguide converter followed by a W-band harmonic mixer. By constructing an optical phase-locked loop (OPLL), a high-spectral-purity mm-wave signal with an electrical power of 2.3 µW was successfully generated at 110 GHz with an rms phase fluctuation of 57 mrad. Secondly, the frequency of the mm-wave signal was extended by use of three cascadingly phase-locked semiconductor lasers. This technique uses a semiconductor optical amplifier (SOA) to generate four-wave-mixing (FWM) signals as well as to amplify the input signals. When the three lasers were appropriately tuned, two pairs of FWM signals were nearly degenerated. By phase-locking the offset frequency in one of the nearly degenerated pairs, the frequency separations among the three lasers were kept at a ratio of 1:2. Thus, we successfully generated high-purity millimeter-wave optical-beat signals at frequencies at 330.566 GHz with an rms phase fluctuation of 0.38 rad. A detailed analysis of the phase fluctuations was carried out on the basis of measured power spectral densities. The possibility of extending the mm-wave frequency up to 1 THz by using four cascadingly phase-locked lasers was also discussed.