• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2004년도 춘계학술대회 학술발표 논문집 제14권 제1호
• /
• pp.470-473
• /
• 2004

In this paper, the comparison of speech feature parameters for emotion recognition is studied for emotion recognition using speech signal. For this purpose, a corpus of emotional speech data recorded and classified according to the emotion using the subjective evaluation were used to make statical feature vectors such as average, standard deviation and maximum value of pitch and energy. MFCC parameters and their derivatives with or without cepstral mean subfraction are also used to evaluate the performance of the conventional pattern matching algorithms. Pitch and energy Parameters were used as a Prosodic information and MFCC Parameters were used as phonetic information. In this paper, In the Experiments, the vector quantization based emotion recognition system is used for speaker and context independent emotion recognition. Experimental results showed that vector quantization based emotion recognizer using MFCC parameters showed better performance than that using the Pitch and energy parameters. The vector quantization based emotion recognizer achieved recognition rates of 73.3％ for the speaker and context independent classification.

• 말소리와 음성과학
• /
• 제6권3호
• /
• pp.23-32
• /
• 2014

This study presents the prosodic characteristics of 'Motherese' and 'Teacherese (child care teacher and kindergarten teacher)'. 21 mothers and 24 teachers spoke to children in the child care center or kindergarten. Children are in their 4;00-6;11. Speech and articulation rate, number of accentual phrases (APs), number of intonational phrases (IPs), pitch-related factors (f0, pitch range, f0 standard deviation), and intonation slope (mean Absolute, f0, q-tone slope) were measured. 2 groups spoke 2 sentential types (interrogative_ alternative question, declarative_ coordinated sentence) in 2 situations (one accompanied with the children, the other done without children, but pretending as if they were in front of the children). The results indicate that teachers show more noticeable prosodic characteristics than mothers do.

• 말소리와 음성과학
• /
• 제9권4호
• /
• pp.107-114
• /
• 2017

The purpose of the current study was to compare selected acoustic and electroglottographic measures (closed quotient, pitch, and loudness) among vowel phonation, traditional voiced lip trill ($VLT_T$), modified voiced lip trill methods ($VLT_M$). A total of 21 participants without voice complaints produced 4-second long samples using each phonation method. Results indicated that mean closed quotient of $VLT_M$ was higher than that of vowel phonation and $VLT_T$, while its range and standard deviation measures were higher than those of vowel phonation. Mean, range, standard deviation, maximum of pitch measures of $VLT_M$ were higher than those of vowel phonation. Lastly, mean and maximum loudness of the $VLT_M$ were higher than $VLT_T$. In conclusion, the current data indicate the possibility to use the $VLT_M$ as a training method for singing or a strategy to facilitate generalization effect of voice therapy. Current results also reflect the necessity for further study pertaining to the long-term effect of the $VLT_M$ training method. Clinical implications are discussed.

• 대한토목학회논문집
• /
• 제29권5A호
• /
• pp.421-431
• /
• 2009

본 연구는 유도초음파를 기반으로 한 구조물 건전성 모니터링 알고리즘 검증에 관한 것이다. 유도초음파 중 램파를 기반으로 하는 능동검사시스템을 사용하였으며 본 알고리즘의 적용에 관한 기초연구로서 강판에 표면결함인 원형노치를 감지하는 실험을 수행하였다. Pitch-catch방법으로 강판 표면 원형노치에 대한 손상 탐상을 수행하여 손상전후의 $S_0$모드를 비교함으로써 손상 특성을 파악하며 손상전후 신호의 수학적 차에 의해 구한 산란램파(scatter)에 의해 손상을 평가할 수 있다. 손상전후 신호의 $S_0$모드 램파를 분석하기 위해서 연속 가보 웨이블릿 변환을 사용하여 전파시간을 구할 수 있다. 본 알고리즘은 손상모니터링 다각형과 Pitch-catch 방법을 기초로 하며 손상 위치표정값들의 평균과 표준편차에 의해 손상위치표정 및 크기를 정확히 추정 가능함이 검증되었다.

• 대한전자공학회논문지SP
• /
• 제40권5호
• /
• pp.371-377
• /
• 2003

본 논문에서 음성신호를 사용하여 인간의 감정를 인식하기 위한 특징 파라메터 비교에 관하여 연구하였다. 이를 위하여 여러 가지 감정 상태에 따라 분류된 한국어 음성 데이터 베이스를 이용하여 얻어진 음성 신호의 피치와 에너지의 평균, 표준편차와 최대 값 등 통계적인 정보 나타내는 파라메터와 음소의 특성을 나타내는 MFCC 파라메터가 사용되었다. 파라메터들의 성능을 평가하기 위하여 문장 및 화자 독립 감정 인식 시스템을 구현하여 인식 실험을 수행하였다. 성능 평가를 위한 실험에서는 운율적 특징으로 피치와 에너지와 각각의 미분 값을 사용하였고, 음소의 특성을 나타내는 특징으로 MFCC와 그 미분 값을 사용하였다. 벡터 양자화 방법을 사용한 화자 및 문장 독립 인식 시스템을 사용한 실험 결과에서 MFCC와 델타 MFCC를 사용한 경우가 피치와 에너지를 사용한 방법보다 우수한 성능을 나타내었다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1988년도 전기.전자공학 학술대회 논문집
• /
• pp.25-28
• /
• 1988

A speech that play a part of important mediation in the man's conversation is the sound of representation to man's emotion and thought, then voice sound could be verified and identified a speaker's speech by individual property. This study indicates as distribution of pitch in searching for sample number of each pitch with eye in the sound waveform of speaker. We propose the algorithm that judge speaker's emotion state, personality, regional group, age, sex distinction, e.t.c., according to the deviation degree.

• 대한음성학회지:말소리
• /
• 제46호
• /
• pp.103-115
• /
• 2003

In this paper, we proposed Background Model Set algorithm used in the speaker verification to improve calculating confidence measure(CM) in speech recognition. CM is to display relative likelihood between recognized models and antiphone models. In previous method calculating of CM, we calculated probability and standard deviation using all phonemes in composition of antiphone models. At this process, antiphone CM brought bad recognition result. Also, recognition time increases. In order to solve this problem, we studied about method to reconstitute average and standard deviation using BMS algorithm in CM calculation.

• 한국해양공학회지
• /
• 제31권2호
• /
• pp.81-90
• /
• 2017

The 1MW OTEC (Ocean Thermal Energy Conversion) platform was designed for application in equatorial seas. In this study, the OTEC platform was investigated using numerical and experimental methods. An octagon-shaped OTEC platform was investigated using the Ocean Engineering Basin of KRISO. These experiments included various tests of regular waves, irregular waves and irregular waves with current (wave+current). The responses of the platform in regular waves showed good agreement between the numerical and experimental results, including the motion RAO, wave run up, and mean drift force. The peak period of heave and pitch motions were observed around 0.5 rad/s, and the effect of the total reflection was found under short wave conditions. The standard deviation (STD) of the platform motion was checked in irregular waves of equatorial and Hawaiian seas. The STD of the pitch was less than $4^{\circ}$ different from the operability requirement under equatorial conditions and the surge STD of the wave frequency showed good agreement between the numerical and experimental results. The STD values of the surge and pitch were increased 66.6% and 92.8% by the current effects in irregular waves, but the pitch STD was less than $4^{\circ}$ under equatorial conditions. This study showed that the STD of the surge was affected by spring effects. Thus, the watch circle of the platform and tension of the mooring lines must be evaluated for a specific design in the future.

• 대한기계학회논문집A
• /
• 제28권9호
• /
• pp.1408-1414
• /
• 2004

The use of flip-chip technology has many advantages over other approaches for high-density electronic packaging. ACF (anisotropic conductive film) is one of the major flip-chip technologies, which has short chip-to-chip interconnection length, high productivity, and miniaturization of package. In this study, thermal fatigue lift of ACF bonding flip-chip package has been predicted. Elastic and thermal properties of ACF were measured by using DMA and TMA. Temperature dependent nonlinear hi-thermal analysis was conducted and the result was compared with Moire interferometer experiment. Calculated displacement field was well matched with experimental result. Thermal fatigue analysis was also conducted. The maximum shear strain occurs at the outmost located bump. Shear stress-strain curve was obtained to calculate fatigue life. Fatigue model for electronic adhesives was used to predict thermal fatigue life of ACF bonding flip-chip packaging. DOE (Design of Experiment) technique was used to find important design factors. The results show that PCB CTE (Coefficient of Thermal Expansion) and elastic modulus of ACF material are important material parameters. And as important design parameters, chip width, bump pitch and bump width were chose. 2$^{nd}$ DOE was conducted to obtain RSM equation far the choose 3 design parameter. The coefficient of determination ($R^2$) for the calculated RSM equation is 0.99934. Optimum design is conducted using the RSM equation. MMFD (Modified Method for feasible Direction) algorithm is used to optimum design. The optimum value for chip width, bump pitch and bump width were 7.87mm, 430$\mu$m, and 78$\mu$m, respectively. Approximately, 1400 cycles have been expected under optimum conditions. Reliability analysis was conducted to find out guideline for control range of design parameter. Sigma value was calculated with changing standard deviation of design variable. To acquire 6 sigma level thermal fatigue reliability, the Std. Deviation of design parameter should be controlled within 3% of average value.

• 핵의학기술
• /
• 제17권2호
• /
• pp.67-71
• /
• 2013

PET/CT검사에서 CT촬영조건의 변화는 영상의 화질 및 환자가 받는 피폭선량에 영향을 미친다. 본 연구는 CT 매개 변수 중 Pitch와 X-선관 회전시간 변화에 따른 선량대비 CT 영상의 질과 이로 인해 PET상에서 SUV에 미치는 영향을 비교 평가하고자 하였다. Discovery STe PET/CT 장비를 이용하여 영상을 획득하였다. QA Phantom과 AAPM Phantom을 이용한 CT 영상 획득 시 Pitch는 0.562, 0.938, 1.375, 1.75:1로 4단계, X-선관 회전시간은 0.5에서 1.0까지 0.1초씩 증가시켜 6단계로 나누어 총 24개 조합을 적용한 영상을 각각 획득하였다. PET 영상은 $^{18}F-FDG$ 5.3 kBq/mL가 채워진 1994 NEMA PET Phantom을 이용하여 프레임당 2분 30초의 방출영상을 획득하였다. 각 조합의 CT 영상에 관심영역을 설정하고 CT number의 표준편차를 측정하였다. 동일한 영상에서 DLP변화에 따른 영상잡음의 예측값을 계산하여 예측값 대비 실측값의 비율을 구해 선량대비 영상잡음 효과를 비교하는 척도로 사용하였다. AAPM Phantom 영상에서 1.0 mm까지 식별이 가능한 지 확인하였다. NEMA PET Phantom의 방출영상에 관심영 역을 설정하고 SUV를 비교 평가하였다. Pitch가 0.562, 0.938, 1.375, 1.75:1로 변화할 때 영상잡음 효과는 QA Phantom에서 1.00, 1.03, 1.01, 0.96, AAPM Phantom에서 1.00, 1.04, 1.02, 0.97로 측정되었다. 회전시간의 증가에 따른 경우 QA Phantom에서 0.99, 1.02, 1.00, 1.00, 0.99, 0.99이었고, AAPM Phantom에서 1.01, 1.01, 0.99, 1.01, 1.01, 1.01로 SPSS Ver. 18을 이용하여 상관관계를 분석한 결과 피어슨 상관계수는 -0.059로 나타났다. 공간분해능에 대한 평가는 24개의 조합 모두에서 1.0 mm까지 육안으로 구별이 가능하였다. SUV의 경우 평균 SUV는 모든 조합에서 1.1로 모두 동일한 값을 나타내었다. Pitch 변화에 따른 CT 영상 평가에서 1.75:1을 적용 시 선량대비 가장 적은 영상잡음 효과를 보이며 공간분해능과 SUV에는 영향을 미치지 않는다. 그러나 회전시간 변화가 영상에 미치는 영향에는 유의한 차이가 없음을 알 수 있다. 결과에서와 같이 각 장비에 따른 선량대비 영상잡음이 적은 Pitch를 사용하고 환자의 체격에 따른 적절한 X-선관 회전시간을 이용한다면 환자의 피폭선량을 줄이면서 최적의 화질을 얻을 수 있는 프로토콜을 구성하는데 도움이 될 것이라 사료된다.