• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.78-81
• /
• 1999

본 논문에서는 인간의 정보처리 기술을 모방한 신경망과 한국어 음절 구성의 특성을 이용하여 음절을 기반으로 하는 신경망 음성인식 방법을 제안한다. 제안한 방법에서는 임계비율을 정의하여 한국어 음절을 구성하는 초성$\cdot$중성$\cdot$종성을 구분하였고, 구분된 음절의 일부 구간 특징을 학습 및 인식의 특징 패턴으로 사용하여 음성인식 시스템의 전체적인 처리 단계를 줄였다. 한국어 숫자음 인식에 대한 성능 평가에서 20대 남성과 여성을 대상으로 화자 종속에서 $96.5\%$의 인식률을 화자 독립에서 $93\%$의 인식률을 얻었다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.05a
• /
• pp.649-652
• /
• 2019

최근에 카메라를 사용하여 물체를 인식하고 수량을 측정하는 연구가 진행되고 있다. 하지만 건축 자재와 같이 인접한 동일 물체의 경우, 인식과 측정이 어려우며, 이와 관련된 연구는 부족한 실정이다. 본 논문에서는 이러한 문제를 해결하기 위해 딥러닝과 드론을 사용하는 방법을 통하여 자재 더미의 윗면과 측면의 촬영, 드론 로그를 사용하한 드론 높이와 Yaw값 획득, 영상 내 자재 종류와 영역 정보 등 미리 학습된 딥러닝 모델을 사용한 획득방법, 드론 촬영 시의 Yaw값 차이를 이용해 윗면과 측면이 공통으로 가지는 직선 검색, 자재 더미의 가로, 세로, 높이의 비율 등을 통한 각 화소 비율과 자재 개수의 영역별 가상의 체적값 분석, 드론의 화각, 높이를 사용한 체적 사진측량값 획득, 가상 체적과 사진측량법 사이의 오차 최소화 비율 등을 연구하였다.

• Korean journal of food and cookery science
• /
• v.19 no.4
• /
• pp.529-538
• /
• 2003

The objectives of this study were to measure customers' cognition and overall satisfaction, and to identify relatively important attributes for the overall satisfaction, of home-delivered meals. Questionnaires were distributed to 243 customers. The statistical data analyses were completed by x$^2$-tests, ANOV A, factor analysis, reliability analysis and regression analysis using SPSS version 10. 56.6% of customers get obtained information from the internet, with 31.3% of these using this method at least once a week, but 72.9% of customers used this method less than once per years. The major reasons for ordering home-delivered meals were tired of cooking, more economical and no time to cook. The results were significantly different in relation to age, occupation and monthly income. The major reasons for hesitation about ordering home-delivered meals were meals should be prepared in households, not sanitary and the use of too many artificial flavors. The results for this factor were significantly different in relation to gender, age and monthly income(p<0.01). The most preferred kinds of home-delivery meals were Korean soup (guk), stew, soup (tang), speciality dishes and party dishes. The customer's cognition of kindness of the delivery staff was highest, with food temperature being the lowest among the options. The food and service level factors were derived from a factor based analysis of customer's cognition towards home-delivered meals. The customer's cognition of food taste, food quantity, kindness of delivery staff and packaging container shape were significantly different according to the use frequency and use period. The packaging method, sanitation, kindness of delivery staff, price and taste were the most relatively important attributes for overall satisfaction with home-delivered meals.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.41 no.8
• /
• pp.1182-1189
• /
• 2012

The purpose of this study was to analyze university student' perception of the price of coffee shops using price sensitivity measurements. The questionnaires were distributed to 640 university students in Changwon, Korea from March 13 to April 8, 2011. A total 571 were used for the final analysis, which excluded improperly-completed questionnaires. The results of this study were as follows. Indifference prices (IDPs) were \2,200 for on-campus coffee shops and \2,850 for off-campus coffee shops. The optimum price points (OPP) were \2,300 for on-campus coffee shops and \2,850 for off-campus coffee shops. The price stress range (PSR) was \100 (\2,200~2,300) for on-campus coffee shops and \0 (\2,850~2,850) for off-campus coffee shops. The point of marginal cheapness (PMC) was \1,100 for on-campus coffee shops and \1,500 for off-campus coffee shops, whereas the point of marginal expensiveness (PME) was \3,600 for on-campus coffee shops and \4,100 for off-campus coffee shops. The ranges of acceptable prices (RAP) were \2,500 (\1,100~3,600) for on-campus coffee shops and \2,600 (\1,500~4,100) for off-campus coffee shops. On the basis of IDP percentage, university students were more sensitive to off-campus coffee shops than on-campus coffee shops. In contrast, on the basis of PSR and RAP, university students were more sensitive to on-campus coffee shops than off-campus coffee shops. Therefore, the price increase for an americano should be recommended to be from \1,100 to \3,600 for on-campus coffee shops and from \1,500 to \4,100 for off-campus coffee shops. The manager of on-campus coffee shops should be careful when increasing the prices, and the manager of off-campus coffee shops should make it possible for customers to perceive service quality.

• Communications of Mathematical Education
• /
• v.29 no.1
• /
• pp.51-72
• /
• 2015

This study aimed to analyze the characteristic of undergraduate students' perception and preference for mathematics. For this purpose, I surveyed 124 undergraduate students' metaphorical expressions about mathematics. I classified the expressions as four categories: a positive form, a negative form, a mixed form, an undecidable form. I investigated the proportion and characteristic of the metaphorical expressions according to the above four categories. Also, I surveyed the students' preference and nonpreference moments for mathematics and categorized them into 6-cases: elementary school, middle school, high school, university, always, and none. In addition, I examined the students' preference and nonpreference reasons for mathematics and classified them according to the 5-factors: grade factor, affective factor, content factor, teacher factor, and other factors. The results of this study as follows: First, the 27% of university students expressed their metaphorical expressions for mathematics as a positive form, 42% as a negative form, and 27% as a mixed form. Also, the preference rate for mathematics was higher as their school years increase and the main reasons of preference were grade and affective factors. The result of nonpreference rate was also higher as their school year increased. Students said that the contents and grade factor were the main factors among the 5-factors.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.22 no.3
• /
• pp.25-55
• /
• 2011

These results of surveying public librarians can contribute to clearly identifying the status of CHI(consumer health information) services in Korea and to exploring future developments in CHI services. In this study, we investigated whether or not currently Korean public libraries are offering CHI services, whether or not they provide CHI reference services and the frequency of their use, and analyzed librarians' recognition of the need for Korean public libraries to provide the CHI services. The findings are as follows: 1) 70% of Korean public libraries are equipped with health-related books and medical journals, but the rate of libraries that are building CHI-related websites was very low at 5%. 2) The rate of providing the CHI reference services was only 30% of the surveyed public libraries. 3) 82.1% of public libraries did not provide educational programs related to CHI. 4) 82.1% of public libraries did not provide CHI-related services for vulnerable groups. 5) The rate of providing the CHI-related outreach programs and collaborating with CHI-related institutions was no more than about 20% each. These research findings can induce public librarians to provide CHI services in public libraries, and will contribute to activate the CHI-related research in Korea.

• Journal of the Korean Chemical Society
• /
• v.55 no.1
• /
• pp.104-111
• /
• 2011

In this study, high school students' perceptions and attitudes toward nanotechnology are examined through questionnaires gathered from 1704 high school students (five general high schools and two science high schools). As a result, the study shows that high school students have generally high perceptions and positive attitudes toward nanotechnology. Science high school students show more positive attitudes and correct knowledge toward nanotechnology rather than general high school students do. High school students get information on nanotechnology via various sources such as TV and internet, whereas they have rarely perceived to get information on nanotechnology in textbooks and science class. They think that newspaper and TV programs are proper methods to introduce new science technology and get information on the advanced science and technology through the internet and science teachers. High school students show positive attitudes toward the application of nanotechnology, whereas they show negative attitudes toward personal information chips and nano-taste enhancer. And they trust nanotechnology researchers and science teachers, whereas they rarely trust about government, public organizations, and internet.

• Proceedings of the KIEE Conference
• /
• 2003.07d
• /
• pp.2516-2518
• /
• 2003

본 논문은 고차 뉴런의 문제점으로 지적되고 있는 뉴런이 방대하게 증가하는 문제를 해결하고자, 최적의 뉴런을 생성하고 생성되어진 고차 뉴런 중 일정 비율로 뉴런의 연결강도를 도태시켜 감에 따라 네트워크상에 나타나는 특성을 비교하였다. 본 논문은 고차 뉴런을 이용한 Kohonen의 자기 조직화 맵의 고차 뉴런부에 일정 비율로 연결강도를 도태한 후 인식률을 얻는 형태로 시뮬레이션을 하였다. 특히, 종래 형태의 고차 뉴런을 이용한 Kohonen 자기 조직화 맵의 알고리즘을 변형없이 사용하였으며 중복되는 뉴런을 최대한 억제하기 위해 2차 뉴런만을 생성한 네트워크 구조 위에 입력 데이터의 특징을 유지하고 고차 뉴런의 특징을 더욱 활성화하기 위해 일정한 양의 연결강도를 도태시킴으로써 출력면에서 국소집중 반응에 의한 정확한 인식률 향상 등을 조사하는 시뮬레이션을 하였다. 본 제안 모델의 특성을 살펴보기 위해 60개의 데이터로 이루어진 금속 소나 음데이터와 암석 소나 음 데이터를 이용하여 금속인지 암석인지를 판별하는 시뮬레이션을 하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.05a
• /
• pp.540-543
• /
• 2020

본 논문에서는 대표적인 이미지 분류 모델인 CNN(Convolutional Neural Network)과 시간에 따른 이미지의 변화를 학습할 수 있는 LSTM(Long Short-Term Memory) 기반의 온라인 서명인식 모델을 제안한다. 실제로는 위조서명을 미리 구하기 어렵다는 사실을 고려해 서명검증 대상자가 아닌 타인의 진서명과 대상자의 일반 필기 데이터를 음의 데이터로서 학습에 사용하였다. 실험 결과, 전체 이미지 중 서명 부분의 비율에 따라 좋은 성능을 보이는 검증 모델이 다르며 Accuracy 성능지표를 통해 이 비율이 높거나 낮을 경우 CNN-LSTM 이, 중간일 경우 CNN 이 적합하다는 것을 확인하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07b
• /
• pp.856-858
• /
• 2005

음성인식 시스템에서 입력된 음성 데이터에 대해 비인식 대상을 거부하는 기능은 신뢰도 보장 측면에 있어서 상당히 중요하며, 신뢰도를 높이기 위해서는 단순한 인식기능 외에 부적절한 입력 패턴의 거부 기능이 필요하다. 본 논문에서는 이러한 신뢰성 문제를 해결하기 위하여 음소기반 인식 네트워크에서 필러 모델 방법과 단어 검출률 방법을 사용하여 실험하였고, 문장의 단어 수에 따른 두 방법의 문장 거부 성능을 FAR과 FRR의 평균을 최소화 하는 값을 각각 구함으로써 비교${\cdot}$분석 하였다. 그 결과 필러모델 방법이 좀 더 나은 거부 성능을 보였고, 단어 검출률을 이용하는 방법이 인식 네트워크를 전부 거치지 않아도 되므로 실행속도와 메모리 절약에서 효과적이었다.