• 한국멀티미디어학회논문지
• /
• 제6권7호
• /
• pp.1312-1320
• /
• 2003

본 논문에서는 하나의 은닉층을 가지는 다층 구조 신경망이 고려되었다. 다층 구조 신경망에서 널리 사용되는 오루 역전파 학습 방법은 초기 가중치와 불충분한 은닉층 노드 수로 인하여 지역 최소화에 빠질 가능성이 있다. 따라서 본 논문에서는 퍼지 단층 퍼셉트론에 ART1을 결합한 방법으로, 은닉층의 노드를 자가 생성(self-generation)하는 퍼지 지도 학습 알고리즘을 제안한다. 입력층에서 은닉층으로 노드를 생성시키는 방식은 ART1을 수정하여 사용하였고, 가중치 조정은 특정 패턴에 대한 저장 패턴을 수정하도록 하는 winner-take-all 방식을 적용하였다. 제안된 학습 방법의 성능을 평가하기 위하여 학생증 영상을 대상으로 실험한 결과. 기존의 오류 역전파 알고즘보다 연결 가중치들이 지역 최소화에 위치할 가능성이 줄었고 학습 속도 및 정체 현상이 개선되었다.

• 정보처리학회논문지B
• /
• 제14B권3호
• /
• pp.215-222
• /
• 2007

최근 정보기술의 발달과 함께 문자 인식의 중요성이 높아지고 있다. 특히, 유비쿼터스 시대가 도래하면서 개인휴대용 정보 단말기, 태블릿 PC 등 유비쿼터스 컴퓨팅 장비가 급속도로 대중화 되고 있다. 이에 사람마다 다양한 필체로 인한 문제가 발생하고 있으며, 인식률을 높일 수 있는 문자 인식에 대한 연구가 필요한 실정이다. 본 연구에서는 다중 해상도 병합을 이용한 수정된 적응 공명 이론 신경망을 제안한다. 이는 자율 학습 신경망과 다중 해상도의 관점에서 접근하여 문자 인식 문제에 적용시켜 본 것이다. 노이즈와 문자 특성 정보를 구별하고 인식률을 높이기 위해 고해상도와 저해상도 정보를 같이 이용하는 다중 해상도 병합 방법을 제안한다. 또한, 다중 해상도 병합 방법의 효과를 극대화할 수 있는 적응 공명 이론 신경망의 유사도 측정 방법을 제안하여 기존의 방법보다 우수한 실험 결과를 제시하였다.

• 한국의상디자인학회지
• /
• 제24권1호
• /
• pp.1-14
• /
• 2022

In the past 10 years, the luxury goods market growth has significantly outpaced other consumer goods categories. Its market value exceeded US $1.2 trillion globally in 2018. In response to this significant consumer demand, the product portfolio of luxury brands has been modified, as several products that previously were exclusive have now released lower-priced versions. This phenomenon is called new luxury. These new luxury products, however, have caused the democratization of the luxury that threatens the exclusivity of luxury brands. In this circumstance, luxury consumers are becoming disillusioned with the vulgarization of luxury goods and prefer exclusive luxury experiences. In order to keep pace with the changes in the luxury consumer needs and promote the brand image with a creative and luxurious connotation, luxury brands are increasingly associating with contemporary art through diverse channels. However, there is a lack of theoretical research investigating the main points to be considered when applying contemporary art to luxury brands and attempting effective brand communication. Therefore, this study aims to provide insights into the key issues for luxury brands when applying contemporary art for effective brand communication. An in-depth case study with the luxury fashion brand, Louis Vuitton, was carried out to better understand contemporary art applications and obtain analytical insights. Through the research, key insights were identified as follows: 1) setting a clear objective for art application based on an in-depth understanding for both the art and the luxury brand, 2) sending harmonized and coherent messages from all perspectives by using a variety of different communication channels, 3) providing a meaningful customer experience by considering the characteristics of contemporary art.

• 정보처리학회논문지B
• /
• 제9B권1호
• /
• pp.119-128
• /
• 2002

본 논문에서는 차량 영상으로부터 동적인 임계화 방법과 개선된 성능의 학습 알고리즘에 의한 신경망을 이용하여 차량 번호판 인식방법을 제안하였다. 제안된 방법에서 번호판 영역은 차량 영상의 구조적 속성을 이용한 동적인 임계화 방법과 밀집비율을 함께 고려하여 추출하였다. 추출된 영역으로부터의 개별문자와 숫자는 윤곽선 추적 알고리즘을 이용하여 각각 추출하였으며, 그들의 인식을 위해서 수정된 ART1과 지도 학습 방법을 결합한 개선된 성능의 신경망을 이용하였다. 제안된 방법의 성능을 확인하기 위해서 실제 차량 번호판들을 대상으로 실험한 결과, 기존의 그레이 명암이나 RGB 컬러 정보들을 이용하는 방법보다 추출률이 개선되었으며, 인식성능도 기존의 오류 역전파 알고리즘의 신경망보다 우수한 성능이 있음을 확인하였다.

• Restorative Dentistry and Endodontics
• /
• 제48권2호
• /
• pp.19.1-19.12
• /
• 2023

Objectives: The aim of the study was to evaluate the clinical and radiographic success of the Hall technique (HT) and atraumatic restorative treatment (ART) restorations using high-viscosity glass-ionomer cement for the management of occlusal carious lesions in primary molars. Materials and Methods: This randomized clinical study observed 40 children (aged 5-6 years). For each child, one tooth was treated with HT and one with ART. The primary outcome measures for HT restorations were successful, minor, and major failure rates. Clinical evaluations of ART restorations were performed according to the modified United States Public Health Service criteria during 18-month follow-up. McNemar test was used for statistical analysis. Results: Thirty of 40 (75%) participants returned for 18 months of follow-up. In the clinical evaluations of teeth that were treated with HT, the patients did not have complaints of pain or other symptoms, all crowns remained in the oral cavity, the gums were healthy, and the teeth were functional in all evaluations. At the end of the 18-month follow-up, the surface texture and marginal integrity criteria of ART restorations were recorded as 26.7% and 33.3%, respectively. In the radiographic evaluation of 30 patients treated with ART and HT, all restorations were considered successful. Conclusions: The 18-month clinical and radiographic results after treatments applied to single-surface cavities in anxious children showed that both treatment methods were successful.

• 한국지역사회생활과학회지
• /
• 제23권3호
• /
• pp.233-243
• /
• 2012

This study was performed to determine the quality characteristics of Jeungpyun with added modified starch. The addition of modified starch did not change the pH level of the samples, but it did increase the volume. The b-value of Jeungpyun with 3% modified starch was significantly lower than that of the control group. The addition of 3% modified starch also significantly increased the hardness, adhesiveness, gumminess and chewiness of the experimental group over the control group. In the test of sensory characteristics, the cell size of Jeungpyun with 2 and 3% modified starch was larger than that of the control group. Cell uniformity of Jeungpyun with 2% modified starch was lower than that of the control group. The hardness of Jeungpyun with modified starch was significantly higher than that of the control group. In the test of taste preference, Jeungpyun with modified starch beat the control group substantially. The quality of Jeungpyun with 1 and 3% modified starch showed a higher overall score than that of the control group.

• 복식
• /
• 제57권2호
• /
• pp.132-142
• /
• 2007

School uniforms are mass produced by large companies and prominent fashion designers. However many high school students modify their uniform or use fashion accessaries to pursue their individualities and their unique style. The objectives of this research were to identify uniform modification and fashion accessory wearing according to the attitudes toward uniform of high school students. The results were as follows; 1. Factors of the attitudes toward school uniform were style and brand, fashion, role intimation, individuality, and neatness. High school students were classified as the uniform indifference group, the pro-uniform group, and the centrist-uniform group according to their attitudes toward uniform. 2. Many students in academic or boys' high schools were classified into the centrist-uniform stoup, and many in art & physical educational or coeducational high schools were classified into the pro-uniform group. The pro-uniform group modified uniform more and stated that the regulation on the fashion accessory wearing was very strict. The uniform indifference group modified uniform less than other groups and complained less about regulation of fashion accessary wearing. 3. Students in art & physical educational high schools showed more positive tendency toward uniform attitudes, uniform modification, and fashion accessary regulation compliances.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2004년도 심포지엄 논문집 정보 및 제어부문
• /
• pp.95-97
• /
• 2004

This paper estimates Adaptive Resonance Theory 1(ART1) as a vigilance parameter of pattern clustering algorithm. Inherent characteristics of the model are analyzed. In particular the vigilance parameter ${\rho}$ and its role in classification of patterns is examined. Our estimates show that the vigilance parameter as designed originally does not necessarily increase the number of categories with its value but can decrease also. This is against the claim of solving the stability-plasticity dilemma. However, we have proposed a modified vigilance parameter estimate criterion which takes into account the problem of subset and superset patterns and stably categorizes arbitrarily many input patterns in one list presentation when the vigilance parameter is closer to one.

• Journal of Mechanical Science and Technology
• /
• 제14권1호
• /
• pp.1-10
• /
• 2000

This work shows that it is possible to obtain information about more than one parameter from acoustic field information. A variety of ultrasonic Lamb wave modes were utilized to reconstruct thickness and density of an isotropic plate. An image reconstruction of one parameter (thickness of a plate) was carried out for four cases, i.e., the lowest symmetrical and anti symmetrical modes, and the fastest symmetrical and anti symmetrical Lamb waves among multiple modes. For two parameter reconstructions (thickness and density), the image processing was performed using the lowest symmetrical and antisymmetrical modes simultaneously. In this work, a modified version of algebraic reconstruction technique (ART), which is a form of finite-series expansion method, was employed to reconstruct the ultrasonically computed tomographic images. Results from several sample geometries are presented.

• ETRI Journal
• /
• 제35권3호
• /
• pp.542-545
• /
• 2013

State-of-the-art voice codecs have been developed to extend the input bandwidth to enhance quality while maintaining interoperability with a legacy codec. Most of them employ a modified discrete cosine transform (MDCT) for coding their extended band. We propose a source filter model-based coding algorithm of MDCT spectral coefficients, apply it to the ITU-T G.711.1 super wideband (SWB) extension codec, and subjectively test it to validate the model. A subjective test shows a better quality over the standardized SWB codec.