• ETRI Journal
• /
• 제45권4호
• /
• pp.678-689
• /
• 2023

This paper describes a novel end-to-end deep generative model-based speaker recognition system using prosodic features. The usefulness of variational autoencoders (VAE) in learning the speaker-specific prosody representations for the speaker recognition task is examined herein for the first time. The speech signal is first automatically segmented into syllable-like units using vowel onset points (VOP) and energy valleys. Prosodic features, such as the dynamics of duration, energy, and fundamental frequency (F₀), are then extracted at the syllable level and used to train/adapt a speaker-dependent VAE from a universal VAE. The initial comparative studies on VAEs and traditional autoencoders (AE) suggest that the former can efficiently learn speaker representations. Investigations on the impact of gender information in speaker recognition also point out that gender-dependent impostor banks lead to higher accuracies. Finally, the evaluation on the NIST SRE 2010 dataset demonstrates the usefulness of the proposed approach for speaker recognition.

• 방송공학회논문지
• /
• 제19권2호
• /
• pp.166-173
• /
• 2014

최근 다양한 방송 및 영상 분야에서 사람의 행동을 인식하여는 연구들이 많이 이루어지고 있다. 영상은 다양한 형태를 가질 수 있기 때문에 제약된 환경에서 유용한 템플릿 방법들보다 특징점에 기반한 연구들이 실제 사용자 환경에서 더욱 관심을 받고 있다. 특징점 기반의 연구들은 영상에서 움직임이 발생하는 지점들을 찾아내어 이를 3차원 패치들로 생성한다. 이를 이용하여 영상의 움직임을 히스토그램에 기반한 descriptor(서술자)로 표현하고 학습기반의 판별기로 최종적으로 영상내에 존재하는 행동들을 인식하였다. 그러나 단일 판별기로는 다양한 행동을 인식하기에 어려움이 있다. 따라서 이러한 문제를 개선하기 위하여 최근에 다중 판별기를 활용한 연구들이 영상 판별 및 물체 검출 영역에서 사용되고 있다. 따라서 본 논문에서는 행동 인식을 위하여 support vector machine과 sparse representation을 이용한 decision-level fusion 방법을 제안하고자 한다. 제안된 논문의 방법은 영상에서 특징점 기반의 descriptor를 추출하고 이를 각각의 판별기를 통하여 판별 결과들을 획득한다. 이 후 학습단계에서 획득된 가중치를 활용하여 각 결과들을 융합하여 최종 결과를 도출하였다. 본 논문에 실험에서 제안된 방법은 기존의 융합 방법보다 높은 행동 인식 성능을 보여 주었다.

• 대한전기학회논문지:전력기술부문A
• /
• 제53권11호
• /
• pp.596-603
• /
• 2004

This paper proposes an expert system based on the pattern recognition method which can enhance the accuracy and effectiveness of real-time bus reconfiguration strategy for the transfer of faulted load when a main transformer fault occurs in the automated substation. The minimum distance classification method is adopted as the pattern recognition method of expert system. The training pattern set is designed MTr by MTr to minimize the searching time for target load pattern which is similar to the real-time load pattern. But the control pattern set, which is required to determine the corresponding bus reconfiguration strategy to these trained load pattern set is designed as one table by considering the efficiency of knowledge base design because its size is small. The training load pattern generator based on load level and the training load pattern generator based on load profile are designed, which are can reduce the size of each training pattern set from max L/sup (m+f)/ to the size of effective level. Here, L is the number of load level, m and f are the number of main transformers and the number of feeders. The one reduces the number of trained load pattern by setting the sawmiller patterns to a same pattern, the other reduces by considering only load pattern while the given period. And control pattern generator based on exhaustive search method with breadth-limit is designed, which generates the corresponding bus reconfiguration strategy to these trained load pattern set. The inference engine of the expert system and the substation database and knowledge base is implemented in MFC function of Visual C++ Finally, the performance and effectiveness of the proposed expert system is verified by comparing the best-first search solution and pattern recognition solution based on diversity event simulations for typical distribution substation.

• 대한공업교육학회지
• /
• 제38권2호
• /
• pp.218-239
• /
• 2013

본 연구는 남자대학생의 군 의무복무계획과 진로장벽인식의 정도를 파악하고, 군 의무복무계획 수준에 따른 진로장벽인식과 진로준비행동의 차이를 살펴보는데 그 목적이 있다. 이를 위해 대전 충남지역 남자 대학생 284명을 대상으로 군 의무복무계획, 진로장벽인식, 진로준비행동 척도를 사용하여 설문을 실시하고 군 의무복무계획과 진로장벽인식 정도, 군 의무복무계획 수준에 따른 진로장벽인식과 진로준비행동의 차이를 알아보기 위해 기술통계, 상관분석, t검정을 실시하였다. 연구 결과 첫째, 남자대학생의 79.2%가 군 의무복무계획을 가지고 있는 것으로 나타났다. 그러나 각 하위요인을 살펴보면 전체적인 군 의무복무계획의 수준은 높은 편이지만 계획을 세우기 위한 실제적 행동적 노력을 행하는 남자대학생들은 비교적 적다는 것을 알 수 있다. 또한 남자대학생의 73.9%가 진로에 있어 군 의무복무를 장벽으로 인식하고 있으며, '매우 그렇다'고 응답한 비율이 최대 집단(16.9%)으로 나타났다. 또한, 각 문항별 응답분포를 통해 군 의무복무 시기의 문제, 시간에 대한 기회비용과 남들보다 뒤쳐질 수 있다는 어려움을 높게 인식한다는 것을 알 수 있다. 둘째, 군 의무복무계획 상 하위 집단 간 진로장벽인식의 차이는 유의미하게 나타나지 않았으나 하위요인인 중요성의 경우 상 하위 집단 간 진로장벽인식의 유의미한 차이가 나타났다. 다시 말해서, 남자대학생에게 군 의무복무라는 장벽은 군 의무복무계획 수준과 상관없이 자신들의 진로과정에 있어 방해요소로 인식되고 있으며, 군 의무복무계획의 중요성을 높게 인식하고 있는 집단이 그렇지 않은 집단보다 군 의무복무를 더 큰 장벽으로 인식하고 있음을 보여준다. 셋째, 군 의무복무계획의 상 하위 집단 간 진로준비행동의 차이는 유의미하게 나타났다. 즉, 군 의무복무계획 수준이 높은 학생들이 상대적으로 높은 진로준비행동을 보였다는 것은 군 의무복무계획의 필요성을 시사해주는 것으로 볼 수 있다. 본 연구를 통해 남자대학생과 진로상담자에게 군 의무복무계획의 필요성을 제시하였다. 또한 군 의무복무 전 대상자를 대상으로 한 선행연구들을 거의 찾아볼 수 없는 실정에서 한국적 진로발달과정의 독특한 특성인 '군 의무복무'에 관한 경험적 자료를 축적하는데 의의가 있다.

• 한국과학교육학회지
• /
• 제14권2호
• /
• pp.225-235
• /
• 1994

Concidering environmental education as an ultimate resolution for environmental problems, we conducted a study focusing on affective matters. An instrument was developed to evaluate attitudes of elementary and middle school students toward environmental problems. To develop a reliable Likert-type evaluation instrument scale with which emotional intensity could be judged, mean, standard deviation, response frequency distribution, discrimination index, reliability were calculated. As a result, 21 statements for recognition level and 14 statements for behavioral level were made(The Cronbach alpha coefficient of the instrument was .786). This instrument was used to evaluate 5th and 6th grade elementary school students and 1st and 2nd grade middle school students(total number of subjects was 980). The result of this survey can be summarized as follows. 1. Students recognized the seriousness of environmental problems but they did not behave in such a manner as to prevent it. 2. As a result of t-test, behavioral level score of elemenatary school students was significantly higher than that of middle school students(p<.001). 3. This study showed that there was a significant correlation between the recognition level score and the behavioral level score(r=.386, p<.001). 4. Two-Way ANOVA was used to analyze that there was any significant difference according to grade and sex. The results were as follows. (1) No significant difference was found in total score. (2) On recognition level, female students' score was signigicantly higher than that of male students(p<.01). (3) On behavioral level, higher-grade students' score was lower than that of lower-grade-students (p<.001).

• 대한치과기공학회지
• /
• 제42권4호
• /
• pp.379-387
• /
• 2020

Purpose: This study aimed to explore methods to improve the quality recognition and satisfaction level of non-face-to-face lectures among students in the departments of dental technology and dental hygiene. Methods: This survey was conducted to assess the status and preference of non-face-to-face lectures and the level of awareness and satisfaction regarding the quality of these lectures among 179 students of dental technology and 295 students of dental hygiene. Statistical analyses were performed using frequency analysis, independent sample t-test, one-way ANOVA (post-hoc Duncan), Welch analysis (post-hoc Games-Howell), and hierarchical multiple regression analysis. Results: Factors that affected the ability to assess the quality of non-face-to-face lectures were the department, the method of non-face-to-face lectures, the most preferred method for conducting lectures, the level of awareness regarding the quality of face-to-face lecture, and satisfaction level. It has 71.5% explanatory power. Moreover, factors that influenced the satisfaction level of non-face-to-face lectures included the department, grade, the highest satisfied non-face-to-face teaching method, the most effective theoretical non-face-to-face teaching method, the most preferred teaching methods, and the ability to assess quality of face-to-face lectures. It has 46.8% explanatory power. Conclusion: Non-face-to-face classes should be designed and developed for web-based programs to improve the motivation and achievement level of the students and encourage interaction between the professors and students. Our findings suggest that educators should strive to achieve optimal educational effects by efficiently combining face-to-face and non-face-to-face lectures.

• 대한방사선기술학회지:방사선기술과학
• /
• 제41권2호
• /
• pp.163-170
• /
• 2018

The purpose of this study was to analyze the structural relationship between radiation hazards and radiation benefits effecting on overcoming recognition of Fukushima nuclear disaster (FND) in Japan by using structural equation modeling (SEM). The subjects were 248 undergraduates from one university in Chungbuk province in Korea. From June 1, 2017 to July 30, 2017, we conducted a questionnaire survey on the radiation hazards and radiation benefits and on the overcoming recognition of FND. As a result, it showed that the recognition of radiation hazards has a significant effect on the benefits of radiation, but does not directly affect the overcoming recognition of FND. However, the recognition of radiation benefits has been mediating between radiation hazards perception and the overcoming recognition of FND. Therefore, it can be empirically confirmed that despite the radiation hazards the recognition of overcoming the FND depends on the level of radiation benefits by using the SEM.

• 가정과삶의질연구
• /
• 제14권1호
• /
• pp.21-34
• /
• 1996

The purposes of this study are to investigate the degrees of husband's and wife's recognition and the difference between those degrees about the husbands' authoritarian communication to develop the typology of recongnition of husband and wife and also to find group differences caused by the personal variables spousal variables family-environmental variables. The subjects of this research were 328 married couples living in Seoul. The major findings of this study can be summarized as follows: 1) The degrees of husband's and wife's recognition about the husbands' authoritarian communication were not high and there was no significant difference between husband's and wife's recognition. 2) The recognition of husband and wife about the husband's authoritarian communication is classified into 5 types; type of high agreement of both husband and wife(34 married couples) type of middle agreement of both husband and wife(167 married couples) type of low agreement of both husband and wife 26 married couples) type of disagreement with husband's high recognition(51 married couples) type of disagreement with husband's high recognition(51 married couples) are type of disagreement with wife's high recognition(50 married couples) 3) The variables which have significance on the types of recognition of husband and wife about the husband's authoritarian communication were husband's educational level husband's birth order husband's self-esteem husband's alienation husband's sex-role attitudes wife's self-esteem wife's alienation the rights of decision-making between couples household income subjective social class authoritarian behavior of father of husband.

• 한국주거학회논문집
• /
• 제20권1호
• /
• pp.91-99
• /
• 2009

With the development of the internet and the spread of 'digital home', apartment houses equipped with the homepages connecting apartment houses as one cyberspace unit have increased. The virtual community with a homepage in apartment complex have improved the effect on management businesses in the side of a apartment manager and increased the importance of the space with a strong tie between residents in the side of residents. But the apartment residents didn't make full use of the virtual community because they have the low level of recognition of it. Therefore, this study attempted to present basic materials for the scheme to revitalize the virtual community. For this purpose, it attempted to find residents' recognition of and participation in the virtual community and demand for the classified contents conducted before this study. In the research method, it conducted the questionnaire research. The results are as follows. 1) First of all, It was found that the apartment residents had the low level of recognition of the virtual community. There was a significant difference according to the degree and the level of using the internet. Some easy-access programs are necessary. 2) The apartment residents took part in the virtual community about 67 percents because there was a lack of the public relation of the virtual community. 3) It was found that there was a significant positive correlation between the level of recognition and participation. 4) An attempt was made to identify the apartment residents' demand for the type of virtual community contents. It was found that most of contents were needed. 'apartment complex introduction and management information' showed the highest score among them. 5) Most respondents are affirmative to the virtual community. Therefore the residential community could be reinforced if the contents of virtual community is developed and well operated in the future.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제31권1호
• /
• pp.1-7
• /
• 2004

본 논문에서는 필기체 숫자인식을 위해서 계층적으로 서로 다른 레벨의 정보를 표현할 수 있는 구조화된 특징들의 추출 방법과 특징들 사이에 의존도를 이용하여 분류하는 베이지안 망을 제안한다. 이러한 계층적 특징들을 추출하기 위해서 레벨 단위로 가버 필터들을 정의하고, FLD(Fisher Linear Discriminant) 척도를 이용하여 최적화된 가버 필터들을 선택한다. 계층적 가버 특징들은 최적화된 가버 특징들을 이용하여 추출되며, 하위 레벨일수록 더욱 국부적인 정보를 표현한다. 추출된 계층적 가버 특징들의 분류성능 향상을 위해서 가버 특징들 사이의 계층적 의존도를 이용하는 베이지안 망을 생성한다. 본 논문에서 제안하는 방법은 naive Bayesian 분류기, k-nearest neighbor 분류기, 그리고 신경망 분류기들과 함께 필기체 숫자인식에 적용되어 계층적 가버 특징들의 효율성과 계층적 의존도를 이용하는 베이지안 망은 분류성능을 향상시킬 수 있다는 것을 보여준다.