http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
진공 플라즈마 용사법을 통해 형성된 NiCoCrAlY 오버레이 코팅의 반복 산화 거동
유연우(Yeon Woo Yoo),남욱희(Uk Hee Nam),박훈관(Hunkwan Park),박영진(Youngjin Park),이성훈(Sunghun Lee),변응선(Eungsun Byon) 한국표면공학회 2019 한국표면공학회지 Vol.52 No.6
MCrAlY overaly coatings are used as oxidation barrier coatings to prevent degradation of the underlying substrate in high temperature and oxidizing environment of the hot section of gas turbines. Therefore, oxidation resistance in high temperature is important property of MCrAlY coatings. Also, coefficients of thermal expansion (CTE) of MCrAlY have middle value of that of Ni-based superalloys and oxides, which have the effect of preventing the delamination of the surface oxides. Cyclic oxidation test is one of the most useful methods for evaluating the high temperature durability of coatings used in gas turbines. In this study, NiCoCrAlY overlay coatings were formed on Inconel 792(IN 792) substrates by vacuum plasma spraying process. Vacuum plasma sprayed NiCoCrAlY coatings and IN 792 susbstrates were exposed to 1000℃ one-hour cyclic oxidation environment. NiCoCrAlY coatings showed lower weight gain in short-term oxidation. In long-term oxidation, IN 792 substrates showed higher weight loss due to delamination of surface oxide but NiCoCrAlY coatings showed lower weight loss. X-ray diffraction (XRD) analysis showed α-Al2O3 and NiCr₂O₄ was formed during the cyclic oxidation test. Through cross-section observation using scanning electron microscopy (SEM) and electron back scatter diffraction (EBSD) analysis, thermally grown oxide (TGO) layer composed of α-Al₂O₃ and NiCr₂O₄ was formed and the thickness of TGO increased during 1000℃ cyclic oxidation test. β phase in upper side of NiCoCrAlY coating was depleted due to oxidation of Al and outer beta depletion zone thickness also increased as the cyclic oxidation time increased.
한국인의 영어 인식을 위한 문맥 종속성 기반 음향모델/발음모델 적응
오유리,김홍국,이연우,이성로,Oh, Yoo-Rhee,Kim, Hong-Kook,Lee, Yeon-Woo,Lee, Seong-Ro 대한음성학회 2008 말소리 Vol.68 No.-
In this paper, we propose a hybrid acoustic and pronunciation model adaptation method based on context dependency for Korean-English speech recognition. The proposed method is performed as follows. First, in order to derive pronunciation variant rules, an n-best phoneme sequence is obtained by phone recognition. Second, we decompose each rule into a context independent (CI) or a context dependent (CD) one. To this end, it is assumed that a different phoneme structure between Korean and English makes CI pronunciation variabilities while coarticulation effects are related to CD pronunciation variabilities. Finally, we perform an acoustic model adaptation and a pronunciation model adaptation for CI and CD pronunciation variabilities, respectively. It is shown from the Korean-English speech recognition experiments that the average word error rate (WER) is decreased by 36.0% when compared to the baseline that does not include any adaptation. In addition, the proposed method has a lower average WER than either the acoustic model adaptation or the pronunciation model adaptation.
한국인 화자에서 새로운 버전의 Acoustic Voice Quality Index 적용
김근효(Geun-Hyo Kim),이연우(Yeon-Yoo Lee),배인호(In-Ho Bae),박희준(Hee-June Park),권순복(Soon-Bok Kwon) 한국언어청각임상학회 2018 Communication Sciences and Disorders Vol.23 No.4
배경 및 목적: 본 연구에서는 한국인 화자에서 정상음성과 병리적 음성을 정량화하는 Acoustic Voice Quality Index (AVQI)의 새로운 회귀식을 제시하여 기존의 회귀식(AVQIv2, AVQIv3)과 비교하고자 하였다. 또한 AVQI 측정치와 청지각적 평가 간의 상관관계도 알아 보고, 정상과 병리적 음성을 감별할 수 있는 최적의 절사값 및 진단 예측력도 알아보았다. 방법: 한국인 화자 2,369명으로부터 모음연장 발성 /아/ 5초와 표준문단 ‘산책’ 읽기를 통해서 녹음하였다. 모음의 안정구간 2초와 읽기샘플 25음절을 연결하여 Praat 스크립트를 적용하여 유성음 분절을 추출하고 AVQI를 측정하였다. 또한 음성장애의 중증도를 평가하기 위해서, 5명의 언어재활사가 두 가지 청지각적 평가(GRBAS의 Grade척도, CAPE-V의 overall severity 척도)를 사용하였다. 결과: 전체 AVQI 측정치는 AVQIv2, AVQIv4-K, AVQIv3 순으로 높았으며, AVQIv3는 나머지 AVQI 측정치들에 비해서 유의미하게 낮았다. 각 버전 모두 정상과 병리적 음성 집단 간에 유의 미한 차이를 보였다. AVQI와 청지각적 평가 간에는 .82 이상의 높은 상관관계를 보였고, AVQIv4-K와 상관관계가 가장 높았다. ROC 분석을 통해서 모든 AVQI는 모두 .94 이상의 AUC, 9.5 이상의 LR(+), .2 이하의 LR(-)를 나타내었다. 논의 및 결론: AVQIv4-K는 기존 AVQI 버전에 상응하는 높은 수준의 진단적 성능을 가지는 것으로 여겨진다. Objectives: The purpose of the present study was to compute the new version of Acoustic Voice Quality Index (AVQI) and compare with existing AVQI versions. Additionally, we also investigated the correlation between AVQI measurements and auditory-perceptual ratings, and confirmed the optimal cutoff value and diagnostic predictive power to distinguish between normal and pathological voice. Methods: Voice samples of 2,369 native Korean subjects with normal voices (N= 143) and with dysphonia (N= 2,226) were recorded through both 5 seconds of sustained vowel /a/ and connected speech (Korean text “Walk”). We used the Praat script to extract the voiced segments and measure the AVQI. Additionally, two auditory-perceptual ratings (grade in GRBAS and overall severity in CAPE-V) were used by five speech language pathologists to assess the voice disorder severity. Results: Overall AVQI measurements were higher in order of AVQIv2, AVQIv4-K and AVQIv3; and AVQIv3 was significantly lower than other AVQI measures. In each version, there was a significant difference between normal and pathological voice groups. The correlation between AVQI and auditory evaluation was above .82, and the correlation with AVQIv4-K was the highest. All versions of the AVQI showed an above .94 for AUC, above 9.5 for LR(+), and below .2 for LR(–). Conclusion: AVQIv4-K is considered to have a high level of diagnostic power comparable to the existing AVQI versions.
타언어권 화자 음성 인식을 위한 혼잡도에 기반한 다중발음사전의 최적화 기법
김민아,오유리,김홍국,이연우,조성의,이성로,Kim, Min-A,Oh, Yoo-Rhee,Kim, Hong-Kook,Lee, Yeon-Woo,Cho, Sung-Eui,Lee, Seong-Ro 대한음성학회 2008 말소리 Vol.65 No.-
In this paper, we propose a method for optimizing a multiple pronunciation dictionary used for modeling pronunciation variations of non-native speech. The proposed method removes some confusable pronunciation variants in the dictionary, resulting in a reduced dictionary size and less decoding time for automatic speech recognition (ASR). To this end, a confusability measure is first defined based on the Levenshtein distance between two different pronunciation variants. Then, the number of phonemes for each pronunciation variant is incorporated into the confusability measure to compensate for ASR errors due to words of a shorter length. We investigate the effect of the proposed method on ASR performance, where Korean is selected as the target language and Korean utterances spoken by Chinese native speakers are considered as non-native speech. It is shown from the experiments that an ASR system using the multiple pronunciation dictionary optimized by the proposed method can provide a relative average word error rate reduction of 6.25%, with 11.67% less ASR decoding time, as compared with that using a multiple pronunciation dictionary without the optimization.