• Journal of Digital Convergence
• /
• v.10 no.6
• /
• pp.297-302
• /
• 2012

In this paper, we present a object detection and tracking method based on image context analysis. It is robust from the image variations such as complicated background, dynamic movement of the object. Image context analysis is carried out using the hybrid network of k-means and RBF. The proposed object detection employs context-driven adaptive Bayesian framework to relive the effect due to uneven object images. The proposed method used feature vector generator using 2D Haar wavelet transform and the Bayesian discriminant method in order to enhance the speed of learning. The system took less time to learn, and learning in a wide variety of data showed consistent results. After we developed the proposed method was applied to real-world environment. As a result, in the case of the object to detect pass outside expected area or other changes in the uncertain reaction showed that stable. The experimental results show that the proposed approach can achieve superior performance using various data sets to previously methods.

• Proceedings of the Costume Culture Conference
• /
• 2003.04a
• /
• pp.72-73
• /
• 2003

본 연구는 간호사복의 색상 및 문양에 따른 이미지를 분석하고, 색상과 문양이 조합을 이루었을 때 어떤 이미지로 통합되는지를 규명하고, 피험자에 따라 이미지지각의 차이를 알아보는데 목적이 있었다. 연구방법으로는 준실험 연구방법으로 피험자간 설계를 하였으며, 피험자는 대전, 충남지역의 대학생을 대상으로 시행하여 통계에 적합한 739부를 사용하였다. 연구에 사용된 자극물은 간호사복 catalog에서 선택한 모델에게, 색상(흰색, 분홍색, 하늘색, 녹색, 베이지색)과 문양(무지, 줄, 꽃)을 조합한 총 13개의 자극물을CAD simulation으로 제작했다. (중략)

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.27 no.1
• /
• pp.15-26
• /
• 2023

The purpose of this study is to determine the artificial intelligence-based degradation index from the image of the cross-section of the microstructure taken with a scanning electron microscope of the specimen obtained by the creep test of DA-5161 SX, a nickel-based superalloy used as a material for high-temperature parts. It proposes a new method of quantification and proposes a model that predicts degradation based on Bayesian inference without destroying components of high-temperature parts of operating equipment and a creep life prediction model that predicts Larson-Miller Parameter (LMP). It is proposed that the new degradation indexing method that infers a consistent representative value from a small amount of images based on the geometrical characteristics of the gamma prime phase, a nickel-base superalloy microstructure, and the prediction method of degradation index and LMP with information on the environmental conditions of the material without destroying high-temperature parts.

• Proceedings of the Korean Information Science Society Conference
• /
• 2008.06c
• /
• pp.425-428
• /
• 2008

모바일기기에 탑재되어있는 디지털 카메라의 성능이 향상됨에 따라 이를 이용한 사진의 촬영 및 수집이 용이해졌으며, 따라서 사용자 로그정보를 이용하여 방대한 양의 사진을 분석하거나 브라우징해주는 방법들이 연구되고 있다. 본 논문에서는 모바일기기의 불확실한 로그정보와 사진 주석정보를 베이지안 네트워크로 모델링하여 사용자가 겪은 이벤트들을 추론하고 사용자의 일과를 요약해주는 방법을 제안한다. 우선 사진들을 시간과 위치정보에 따라 분할하여 사진그룹목록을 생성하고, 이를 모바일기기에 입력되어있는 사용자의 일정목록과 합하여 임시이벤트목록을 생성한다. 그 뒤 베이지안 네트워크를 이용하여 각 이벤트를 인식하고 이를 가장 잘 나타내는 사진을 선택한다. 제안하는 방법은 선택된 사진들을 나열하여 사진다이어리형식으로 사용자의 일과를 요약하여주며, 이때 특정 이벤트와 매치되는 사진이 없을 경우 미리 정의되어있는 만화 컷을 대신 사용하여 내용이 매끄럽게 이어지도록 하였다.

• Journal of the Korean Society of Costume
• /
• v.62 no.6
• /
• pp.19-37
• /
• 2012

This study was done to define the color range, images and color characteristics of beige in fashion by analyzing the characteristics of beige that appears in contemporary fashion. In reference research, the general characteristics, the color name and range of beige was examined. In investigation research, color characteristics and images of beige that appear in contemporary fashion was defined, and it was discovered that the cases that the use of beige took more than 50% of the entire in the major collection from S/S season in 2007 to F/W season in 2010 as the subject of color analysis. The result of the study could be summarized as follows: First, the color range of beige consisted of the standard color from 2.5YR to 5YR, and color tone was pale and light gray. Second, the analysis of color beige that appear in contemporary fashion didn't show big differences according to year, season, and region. The color tones consisted of light grayish, pale, light, soft focusing colors from 5YR to 10YR. As for the arrangement of colors, it was arranged with black and it was used with the affiliation of YR, R, Y in many cases. Third, beige monochromatic image appeared soft, plain and classic. The arrangement of the image, modern, feminine, luxurious, gentle, intelligent five types were derived. Arrangements with achromatic colors were expressed in contemporary and sophisticated styles and arrangements with chromatic colors appeared to be soft, feminine and luxurious. This study draws the result to apply the fashion image of beige that was insufficient in other various color researches to design various color aspects by defining the image of beige that appears in contemporary fashion. Based on practical analysis for the color beige, it is evident that beige is an important factor and a powerful influence on fashion images.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.846-849
• /
• 2020

하이퍼파라미터(초매개변수) 최적화란 모델의 학습에 앞서 미리 설정해야 하는 값인 하이퍼파라미터의 최적값을 탐색하는 문제이다. 이때의 최적값은 학습을 끝낸 모델의 성능을 가능한 최대치로 높이게 하는 값이다. 한편, 최근 모바일 장치를 이용한 포지셔닝 데이터의 대량 수집이 가능해지면서 이를 활용하여 위치 기반 서비스(Location-Based Service)를 위한 데이터 분석 및 예측에 관한 연구가 활발히 이루어졌다. 그중 이동 경로를 이미지로 패턴화하여 국소 지역 내에서 다음 위치를 예측하는 CNN 모델에 대해서 하이퍼파라미터 튜닝을 진행하였다. 결과적으로 베이지안 최적화(Bayesian Optimization)를 통해 모델의 성능을 평균 3.7%, 최대 9.5%까지 개선할 수 있음을 확인하였다.

• Geophysics and Geophysical Exploration
• /
• v.23 no.3
• /
• pp.157-167
• /
• 2020

With the recent advancement of computer hardware and the contribution of open source libraries to facilitate access to artificial intelligence technology, the use of machine learning (ML) and deep learning (DL) technologies in various fields of exploration geophysics has increased. In addition, ML researchers have developed complex algorithms to improve the inference accuracy of various tasks such as image, video, voice, and natural language processing, and now they are expanding their interests into the field of automatic machine learning (AutoML). AutoML can be divided into three areas: feature engineering, architecture search, and hyperparameter search. Among them, this paper focuses on hyperparamter search with Bayesian optimization, and applies it to the problem of facies classification using seismic data and well logs. The effectiveness of the Bayesian optimization technique has been demonstrated using Vincent field data by comparing with the results of the random search technique.

• Journal of Korea Society of Industrial Information Systems
• /
• v.25 no.4
• /
• pp.9-16
• /
• 2020

Pronunciation recognition systems that use only video information and ignore voice information can be applied to various customized services. In this paper, we develop a system that applies a Bayesian classifier to distinguish Korean vowels via lip shapes in images. We extract feature vectors from the lip shapes of facial images and apply them to the designed machine learning model. Our experiments show that the system's recognition rate is 94% for the pronunciation of 'A', and the system's average recognition rate is approximately 84%, which is higher than that of the CNN tested for comparison. Our results show that our Bayesian classification method with feature values from lip region landmarks is efficient on a small training set. Therefore, it can be used for application development on limited hardware such as mobile devices.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.33 no.6
• /
• pp.990-1001
• /
• 2009

The purpose of this study was to investigate the effect of perceiver's gender, clothing, and hairstyle on the visual evaluation of men's professionalism and preference. A quasi-experimental method by questionnaire was used. The experimental design was a $2\times8\times2$ (perceiver's gender $\times$ clothing $\times$ hairstyle) factorial design by 3 independent variables. The stimuli were 16 photographs of a man in his twenties. The upper clothing of the man included tailored collar jackets in beige and dark blue colors, and jumpers and sweaters in beige, dark blue, and red colors. The lower clothing of the men included jean pants. Two types of the hairstyles included short hair and medium length hair. The subjects were 208 men and 223 women in Seoul, Korea. Wearing a beige sweater with jean pants was evaluated high in intellectual image, a red jumper was perceived low in intellectual image, and a beige tailored collar jacket was evaluated low in potent image. Men's short hairstyle was evaluated to be more professional than the medium length hair. Male perceivers liked short hair more than medium length hair, but female perceivers evaluated both hairstyles similarly. In the case of women, the preferences of tailored collared jacket and soutien collared jumper were similar, but jumper was preferred to jacket in the case of men. Male perceivers showed more positive feedback towards jean pants with soutien collared jumper than jeans with tailored collared jacket, which indicated that men showed more conservative attitude towards the outfit than women. The man who was wearing a jumper with short hair was evaluated positively and the man who was wearing a jacket with medium length hair was evaluated negatively when the attires were coordinated with jean pants. In conclusion, medium length hairstyle with a beige jacket and short hairstyle with a red sweater were evaluated as professional image; and the results indicated that clothing and hairstyle interact with each other and influence the evaluation of professionalism.

• The KIPS Transactions:PartB
• /
• v.9B no.6
• /
• pp.759-764
• /
• 2002

Although image occupies a large part of importance on the Web documents, there have not been many researches for analyzing and understanding it. Many Web images are used for carrying important information but others are not used for it. In this paper classify the Web images from presently served Web sites to erasable or non-erasable classes. based on machine learning methods. For this research, we have detected 16 special and rich features for Web images and experimented by using the Baysian and decision tree methods. As the results, F-measures of 87.09%, 82.72% were achived for each method and particularly, from the experiments to compare the effects of feature groups, it has proved that the added features on this study are very useful for Web image classification.