• Proceedings of the KSPS conference
• /
• 2003.10a
• /
• pp.145-148
• /
• 2003

To improve the recognition performance of Korean connected digit telephone speech, in this paper, both Aurora feature extraction method that employs noise reduction 2-state Wiener filter and DWFBA method are investigated and used. CMN and MRTCN are applied to static features for channel compensation. Telephone digit speech database released by SITEC is used for recognition experiments with HTK system. Experimental results has shown that Aurora feature is slightly better than MFCC and DWFBA without channel compensation. And when channel compensation is included, Aurora feature is slightly better than DWFBA with MRTCN.

• Journal of the Korean Society for Precision Engineering
• /
• v.18 no.2
• /
• pp.132-139
• /
• 2001

In this research, we have developed a load/unload device capable of correcting the position automatically. Characteristic technologies such as compensation, control, guidance and communication have been modified and implemented on an existing electromagnetic guided AGV, helping to realize open system and distributed cooperation. We have applied the developed AGV with remote control and heterogeneous load/unload mechanisms in a machining system composed of various equipment such as machining centers, CMN and AS/RS and found that the AGV provided position error within $\pm$2mm.

• Phonetics and Speech Sciences
• /
• v.4 no.2
• /
• pp.79-86
• /
• 2012

In this paper, we propose a novel approach to improve histogram equalization for speaker identification. Our method collects all speech features of UBM training data to make a reference distribution. The ranks of the feature vectors are calculated in the sorted list of the collection of the UBM training data and the test data. We use the ranks to perform order-based histogram equalization. The proposed method improves the accuracy of the speaker recognition system with short utterances. We use four kinds of speech databases to evaluate the proposed speaker recognition system and compare the system with cepstral mean normalization (CMN), mean and variance normalization (MVN), and histogram equalization (HEQ). Our system reduced the relative error rate by 33.3% from the baseline system.

• Proceedings of the IEEK Conference
• /
• 2002.06d
• /
• pp.191-194
• /
• 2002

The final purpose of this paper is the enhancement of speech recognition rate under the matched telephone environment between training data and test data. To analyze the effect by the distortion of the changing telephone channel on every call, MFCC is used as the feature parameter and CMN, RTCN, and RASTA are used as channel compensation techniques. For each case, the variation of feature parameters of all phones is analyzed. And, we find recognition rates according to each compensation method using the continuous HMM recognizer, and examine the relationship between variation and recognition rate.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.143-146
• /
• 2002

전화음성의 경우 전화 회선의 채널 대역폭 제한과 통화로 형성시 달라지는 채널의 특성으로 인하여 마이크 음성에 비하여 인식 성능이 많이 저하된다. 본 연구에서는 연속 숫자음 전화음성의 인식율 향상을 위해 채널 왜곡 보상 기법들을 적용하고, HTK 기반의 인식 실험을 통해 보상 기법에 따른 인식 성능을 비교하였다. 채널 왜곡 보상 기법으로 CMN, RASTA, RTCN 등을 적용하고, 각 보상 기법에 따라 HMM의 state 수, mixture 수를 바꾸어 가며 인식 실험한 결과를 제시한다.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.23
• /
• pp.10061-10066
• /
• 2015

Background: The aetiological relationship between human papillomavirus (HPV) infection and cervical cancer (CC) is widely accepted. Our goal was to determine the prevalence of HPV types in Mexican women attending at the Mexican Institute for Social Security from different areas of Mexico. Materials and Methods: DNAs from 2,956 cervical samples were subjected to HPV genotyping: 1,020 samples with normal cytology, 931 with low-grade squamous intraepithelial lesions (LGSIL), 481 with high grade HGSIL and 524 CC. Results: Overall HPV prevalence was 67.1%. A total of 40 HPV types were found; HPV16 was detected in 39.4% of the HPV-positive samples followed by HPV18 at 7.5%, HPV31 at 7.1%, HPV59 at 4.9%, and HPV58 at 3.2%. HPV16 presented the highest prevalence both in women with altered or normal cytology and HPV 18 presented a minor prevalence as reported worldwide. The prevalence ratio (PR) was calculated for the HPV types. The analysis of PR showed that HPV16 presents the highest association with CC, HPV 31, -33, -45, -52 and -58 also demonstrating a high association. Conclusions: The most prevalent HPV types in cervical cancer samples were -16, -18, -31, but it is important to note that we obtained a minor prevalence of HPV18 as reported worldwide, and that HPV58 and -52 also were genotypes with an important prevalence in CC samples. Determination of HPV genotypes is very important in order to evaluate the impact of vaccine introduction and future cervical cancer prevention strategies.

• Journal of the Institute of Convergence Signal Processing
• /
• v.3 no.1
• /
• pp.8-14
• /
• 2002

In this paper, we realized a real-time text-independent speaker recognition system using gaussian mixture model, and applied frame level likelihood normalization method which shows its effects in verification system. The system has three parts as front-end, training, recognition. In front-end part, cepstral mean normalization and silence removal method were applied to consider speaker's speaking variations. In training, gaussian mixture model was used for speaker's acoustic feature modeling, and maximum likelihood estimation was used for GMM parameter optimization. In recognition, likelihood score was calculated with speaker models and test data at frame level. As test sentences, we used text-independent sentences. ETRI 445 and KLE 452 database were used for training and test, and cepstrum coefficient and regressive coefficient were used as feature parameters. The experiment results show that the frame-level likelihood method's recognition result is higher than conventional method's, independently the number of registered speakers.

• Proceedings of the Acoustical Society of Korea Conference
• /
• autumn
• /
• pp.131-134
• /
• 2004

본 논문에서는 GMM(Gaussian Mixture Model)을 이용한 문맥독립 화자 검증 시스템을 구현한 후, arctan 함수를 이용한 정규화 방법을 사용하여 화자검증실험을 수행하였다. 특징파라미터로서는 선형예측방법을 이용한 켑스트럼 계수와 회귀계수를 사용하고 화자의 발성 변이를 고려하여 CMN(Cepstral Mean Normalization)을 적용하였다. 화자모델 생성을 위한 학습단에서는 화자발성의 음향학적 특징을 잘 표현할 수 있는 GMM(Gaussian Mixture Model)을 이용하였고 화자 검증단에서는 ML(Maximum Likelihood)을 이용하여 유사도를 계산하고 기존의 정규화 방법과 arctan 함수를 이용한 방법에 의해 정규화된 점수(score)와 미리 정해진 문턱값과 비교하여 검증하였다. 화자 검증 실험결과, arctan 함수를 부가한 방법이 기존의 방법보다 항상 향상된 EER을 나타냄을 확인할 수 있었다.

• Corrosion Science and Technology
• /
• v.16 no.1
• /
• pp.15-22
• /
• 2017

An oxidation-reduction scheme is an alternative approach for improving the galvanizability of advanced high-strength steel in the continuous hot-dip galvanizing process. Here, we investigated the effect of oxygen partial pressure ($P_{O_2}$) on the oxidation behavior of a transformation-induced plasticity steel containing 1.5 wt% Si and 1.6 wt% Mn during heating to and holding for 60 s at $700^{\circ}C$ under atmospheres with various $P_{O_2}$ values. Irrespective of $P_{O_2}$, a thin amorphous Si-rich layer of Si-Mn-O was formed underneath the Fe oxide scale (a $Fe_2O_3/Fe_3O_4$ bilayer) in the heating stage. In contrast to Si, Mn tended to segregate at the scale surface as $(Fe,Mn)_2O_3$. The multilayered structure of $(Fe,Mn)_2O_3/Fe_2O_3/Fe_3O_4$/amorphous Si-Mn-O remained even after extended oxidizing at $700^{\circ}C$ for 60 s. $Fe_2O_3$ was the dominantly growing oxide phase in the scale. The enhanced growth rate of $Fe_2O_3$ with increasing $P_{O_2}$ resulted in the formation of more Kirkendall voids in the amorphous Si-rich layer and a less Mn segregation at the scale surface. The mechanisms underlying the absence of FeO and the formation of Kirkendall voids are discussed.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.60 no.1
• /
• pp.11-20
• /
• 2018

This study is to calibrate the SWAT (Soil and Water Assessment Tool) water quality of SS (Suspended Solid), T-P (Total Phosphorus), and T-N (Total Nitrogen) by focusing on 2014~2015 drought periods and identify the important parameters. For Gongdo watershed ($366.5km^2$), the SWAT was calibrated for 2 cases of 2002~2006 normal year focusing calibration and 2014~2015 drought focusing calibration respectively. The parameters of N_UPDIS (Nitrogen uptake distribution parameter) and CMN (Rate factor for humus mineralization of active organic nutrients) played important roles for T-N calibration during drought periods. The SWAT SS, T-N, and T-P average $R^2$ (Coefficient of determination) results by focusing on 2014~2015 drought periods calibration showed 0.71, 0.65 and 0.62 while 2002~2006 normal year focusing calibration showed 0.63, 0.58 and 0.50 respectively. Also SWAT SS, T-N, and T-P model efficiency NSE (Nash-Sutcliffe efficiency) results by focusing on drought period (2014~2015) calibrated showed 0.76, 0.77, 0.87 respectively. Even though the SS, T-P parameters were unchanged during the calibration, the SS and T-P results were improved by the hydrological parameters (SCS-CN, SOL_K, SLSOIL) during the drought periods. The SWAT water quality calibration needs to be considered for the movement of SS and nutrients transport especially focusing on the drought characteristics.