• 대한음성학회지:말소리
• /
• 제63호
• /
• pp.125-137
• /
• 2007

Recently, many techniques have been proposed to improve the noise robustness for speaker verification. In this paper, we consider the feature recombination technique in multi-band approach. In the conventional feature recombination for speaker verification, to compute the likelihoods of speaker models or universal background model, whole feature components are used. This computation method is not effective in a view point of multi-band approach. To deal with non-effectiveness of the conventional feature recombination technique, we introduce a subband likelihood computation, and propose a modified feature recombination using subband likelihoods. In decision step of speaker verification system in noise environments, a few very low likelihood scores of a speaker model or universal background model cause speaker verification system to make wrong decision. To overcome this problem, a reliable feature selection method is proposed. The low likelihood scores of unreliable feature are substituted by likelihood scores of the adaptive noise model. In here, this adaptive noise model is estimated by maximum a posteriori adaptation technique using noise features directly obtained from noisy test speech. The proposed method using subband-based reliable feature selection obtains better performance than conventional feature recombination system. The error reduction rate is more than 31 % compared with the feature recombination-based speaker verification system.

• Nuclear Engineering and Technology
• /
• 제55권12호
• /
• pp.4382-4394
• /
• 2023

Passive auto-catalytic recombiners (PARs) are widely used to mitigate a hydrogen hazard. The first step to evaluate the hydrogen safety by PARs is to obtain qualified test data of the PARs for validation of their analytical model. SPARC PAR tests SP8 and SP9 were conducted to evaluate the hydrogen recombination characteristics of a honeycomb-shaped catalyst PAR. To obtain the hydrogen recombination rate from the PAR test data, two methods, Method-1 and Method-2, introduced by the THAI project, were applied. Since a large gradient of hydrogen concentration developed during hydrogen injection can cause a large error in the hydrogen mass obtained by integrating the measured hydrogen concentrations, a gate was installed at the PAR inlet to homogenize hydrogen in the test vessel before the PAR operation in the tests. A computational fluid dynamics (CFD) code with a PAR model was also applied to evaluate the characteristics of the PAR recombination according to the PAR inlet conditions, and the results were compared with those from Method-1 and Method-2. It was confirmed that the recombination rates from Method-1 require a correction factor to be compatible with results from Method-2 and the CFD simulation in the case of the SPARC-PAR tests.

• Current Photovoltaic Research
• /
• 제9권1호
• /
• pp.6-10
• /
• 2021

The mesh mask screen, which is generally used for screen printing metallization of silicon solar cell, requires high squeegee pressure and low printing speed. These requirements are acting as a limiting factor in production yield in photovoltaic industries. In order to improve the productivity, a metal mask, which has high durability and high printing speed, has been researched. In this paper, the characteristics of each solar cell, in which electrodes were formed by using a metal mask and a mesh mask, were analyzed through recombination current density. In particular, the metal-induced recombination current density (Jom) representing the recombination of the emitter-metal interface was calculated using the shading method, and the resulting efficiency and open-circuit voltage were analyzed through the diode equation. As a result of analyzing the proportion of the metal-induced recombination current density to the total emitter recombination current density, it was analyzed that the reduction of the metal-induced recombination current density through the metal mask is an important factor in reducing the total recombination current density of the solar cell.

• 한국의학물리학회지:의학물리
• /
• 제31권4호
• /
• pp.145-152
• /
• 2020

Purpose: In ionization-chamber dosimetry for high-dose-rate electron beams-above 20 mGy/pulse-the ion-recombination correction methods recommended by the International Atomic Energy Agency (IAEA) and the American Association of Physicists in Medicine (AAPM) are not appropriate, because they overestimate the correction factor. In this study, we suggest a practical ion-recombination correction method, based on Boag's improved model, and apply it to reference dosimetry for electron beams of about 100 mGy/pulse generated from an electron linear accelerator (LINAC). Methods: This study employed a theoretical model of the ion-collection efficiency developed by Boag and physical parameters used by Laitano et al. We recalculated the ion-recombination correction factors using two-voltage analysis and obtained an empirical fitting formula to represent the results. Next, we compared the calculated correction factors with published results for the same calculation conditions. Additionally, we performed dosimetry for electron beams from a 6 MeV electron LINAC using an Advanced Markus^® ionization chamber to determine the reference dose in water at the source-to-surface distance (SSD)=100 cm, using the correction factors obtained in this study. Results: The values of the correction factors obtained in this work are in good agreement with the published data. The measured dose-per-pulse for electron beams at the depth of maximum dose for SSD=100 cm was 115 mGy/pulse, with a standard uncertainty of 2.4%. In contrast, the k_s values determined using the IAEA and AAPM methods are, respectively, 8.9% and 8.2% higher than our results. Conclusions: The new method based on Boag's improved model provides a practical method of determining the ion-recombination correction factors for high dose-per-pulse radiation beams up to about 120 mGy/pulse. This method can be applied to electron beams with even higher dose-per-pulse, subject to independent verification.

• 한국음향학회지
• /
• 제28권3호
• /
• pp.279-284
• /
• 2009

화자검증은 발성화자가 제시화자 (claimed speaker)인지 아닌지를 구별하는 것이다. 기존의 화자검증 시스템인 GMM-UBM 방식의 화자검증 시스템은 무잡음 환경에서는 높은 검증성능을 보이지만, 잡음환경에서는 성능이 급격히 떨어지는 단점이 있다. 이런 단점을 극복하기 위해 멀티밴드를 이용한 방법인 특징벡터 재결합방법이 제안되었지만, 특징벡터 재결합방법은 전체 서브밴드 특징벡터들을 사용하여 유사도를 계산하는 단점이 있다. 이런 단점을 극복하기 위해 기 발표된 이전 논문에서 각 서브밴드 유사도를 독립적으로 계산하는 변형된 특징벡터 재결합방법을 제안하였고, 본 논문에서는 변형된 특징벡터 재결합방법과 각 서브밴드들의 신뢰도를 나타내는 신호 대 잡음비를 이용한 가중치를 이용하여 잡음환경에서 기존의 특징벡터 재결합방법에 비해 에러를 28% 감소시켰다.

• Journal of Microbiology and Biotechnology
• /
• 제15권6호
• /
• pp.1346-1352
• /
• 2005

We have developed a modified gene replacement method using PCR products containing short homologous sequences of 40- to 50-nt. The method required $\lambda$-Red recombination functions provided under the control of a temperature-sensitive CI857 repressor expressed from the $P_{lac}$ promoter in the presence of IPTG on an easily curable helper plasmid. The method promoted the targeted gene replacements in the Escherichia coli chromosome after shifting cultures of the recombinogenic host, which carries the helper plasmid, to $42^{\circ}C$ for 15 min. Since this method employs $\lambda$-Red recombination functions expressed from the easily curable helper plasmid, multiple rounds of gene replacements in the E. coli chromosome would be possible. The procedures described herein are expected to be widely used for metabolic engineering of E. coli and other bacteria.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2007년도 한국음성과학회 공동학술대회 발표논문집
• /
• pp.127-130
• /
• 2007

The conventional feature recombination technique is very effective in the band-limited noise condition, but in broad-band noise condition, the conventional feature recombination technique does not produce notable performance improvement compared with the full-band system. To cope with this drawback, we introduce a new technique of sub-band likelihood computation in the feature recombination, and propose a new feature recombination method by using this sub-band likelihood computation. Furthermore, the reliable sub-band selection based on the signal-to-noise ratio is used to improve the performance of this proposed feature recombination. Experimental results shows that the average error reduction rate in various noise condition is more than 27% compared with the conventional full-band speaker identification system.

• 태양에너지
• /
• 제8권2호
• /
• pp.12-18
• /
• 1988

The effect of surface recombination current density on the saturation current density in Si solar cell has been studied. Theoretical model for surface recombination current was set up from emitter transparent model of M.A. Shibib, and saturation current of Si solar cell made by ion implantation method was also measured by digital electrometer. The theoretical surface recombination current density which is the same as saturation surface recombination current density in Shibib model was $10^{-11}[A/cm^2]$ and the measured value was ranged from $8{\times}10^{-10}$ to $2{\times}10^{-9}[A/cm^2]$. Comparing with the ideal p-n junction of Shockley, transparent emitter model shows improved result by $10^2$ order of saturation current density. But there still exists $10^2$ order of difference of saturation current density between theoretical and actual values, which are assumed to be caused by 1) leakage current through solar cell edge, 2) recombination of carriers in the depletion layer, 3) the series resistance effect and 4) the tunneling of carriers between states in the band gap.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1995년도 추계학술대회 논문집 학회본부
• /
• pp.541-543
• /
• 1995

We propose the improved threshold current ratio method to determine the reflectivity of coated facets. The carrier recombination time used in the improved threshold current ratio method depends on the value of facet reflectivities. However, the carrier recombination time used in the conventional threshold current ratio method is constant regardless of facet reflectivities. The difference between the results of the two methods increases as the reflectivity of a coated facet decreases.

• Asian-Australasian Journal of Animal Sciences
• /
• 제14권11호
• /
• pp.1505-1510
• /
• 2001

A method for mapping quantitative trait loci (QTL) was introduced incorporating the information of mixed progeny from complex pedigrees. The method consisted of two steps based on single marker analysis. The first step was to examine the marker-trait association with a mixed model considering common environmental effect and reversed QTL-marker linkage phase. The second step was to estimate QTL effects by a weighted least square analysis. A simulation study indicated that the method incorporating mixed progeny from multiple generations improved the accuracy of QTL detection. The influence of within-genotype variance and recombination rate on QTL analysis was further examined. Detecting a QTL with a large within-genotype variance was more difficult than with a small within-genotype variance. Most of the significant marker-QTL association was detectable when the recombination rate was less than 15%.