• 한국멀티미디어학회논문지
• /
• 제14권11호
• /
• pp.1373-1382
• /
• 2011

본 논문에서는 사용자가 스스로 미소 훈련을 할 수 있도록 자동으로 미소를 분석하는 시스템을 제안한다. 제안된 시스템은 입력 영상으로부터 AdaBoost 알고리즘을 통해 얼굴 영역을 검출한 다음, ASM(active shape model)을 이용하여 생성된 얼굴 형태 모델을 이용하여 얼굴의 특징을 찾는다. 얼굴 특징을 찾은 다음에는 미소 분석에 필요한 입술 라인과 개별 치아 영역을 추출한다. 미소의 정도를 분석하기 위해 입술 라인과 치아와의 관계 판별이 필요한데, 이를 위해 치아 영상에 대해 2차 미분을 실행한 후, 세로축과 가로축에 히스토그램 프로젝션 방법을 이용하여 개별적인 치아 영역을 찾는다. 입술 라인과 개별 치아 영역에 대한 분석을 통해 사용자의 미소 정도를 자동으로 분석하고 결과를 실시간으로 사용자가 직접 확인할 수 있게 해 준다. 본 논문에서 개발된 시스템은 기존에 치과 병원에서 이루어진 미소 훈련을 위한 미소 평가 결과와 8.6% 이하의 오차를 보였으며 사용자가 혼자서도 미소를 훈련하는데 활용할 수 있는 것으로 분석되었다.

• 한국산업정보학회논문지
• /
• 제25권4호
• /
• pp.9-16
• /
• 2020

음성 정보를 배제하고 영상 정보만을 이용한 발음 인식 시스템은 다양한 맞춤형 서비스에 적용될 수 있다. 본 논문에서는 베이지안 분류기를 기반으로 입술 모양을 인식하여 한글 모음을 구분하는 시스템을 개발한다. 얼굴 이미지의 입술 모양에서 특징 벡터를 추출하고 설계된 기계 학습모델을 적용하여 실험한 결과 'ㅏ' 발음의 경우 94%의 인식률을 보였으며, 평균 인식률은 약 84%를 나타내었다. 또한 비교군으로 실험한 CNN 환경에서의 인식률보다 높은 결과를 보였다. 이를 통해서 입술 영역의 랜드 마크로 설계된 특징 값을 사용하는 베이지안 분류 기법이 적은 수의 훈련 데이터에서 보다 효율적일 수 있음을 알 수 있다. 따라서 모바일 디바이스와 같은 제한적 하드웨어에서 응용 가능한 어플리케이션 개발에 활용할 수 있다.

• 대한구순구개열학회지
• /
• 제2권1_2호
• /
• pp.5-14
• /
• 1999

The aim of treatment of cleft lip and palate is to correct the cleft and associated problems surgically and thus hide the anomaly so that patients can lead normal lives. This correction involves surgically producing a face that does not attract attention, a vocal apparatus that permits intelligible speech, and a dentition that allows optimal function and esthetics. In neonatal periods, gross distortion of tissues surrounding the cleft requires considerable effort and time due to post operative functional defect and scarring and induces milk feeding problem, malocclusion of deciduous or permanent dentition, congenital missing teeth, skeletal dysplasia. The occurrence of a cleft deformity is a source of considerable shock to the parents of an afflicted baby, and the most appropriate approach is very important things. Thus we tried to analysis of dental arch, shape and size of deformity in cleft patients. The results were obtained as follows. 1. When the cast measurements of UCLP subjects at first visit it was found that the mean length was 9.29mm at the alveolar cleft width, also that was 11.7mm at the anterior width and 14mm at the posterior cleft width. 2. Comparison of UCLP group at first visit and just lip surgery, it was found that the older group showed a insignificant reduction in the width of the cleft in the alveolar, canine, and tuberosity regions. 3. The maxillary casts of the UCLP group at 6 months differ Significantly from those of the at 3 months in both length and width. but there was no statistical difference except anterior ridge length of nonclefted site. 4. Comparison at 6 months and 18 months, there was a greater change in length of the alveolar cleft width, intercanine width, and anterior cleft width. Maxillary arch became wider at both the canine region and intertuberosity region. also posterior anteroposterior length was increased but anterior AP length was decreased from 8.1mm to 7.7mm. There was meaningful increase at intertuberosity length; however, a significant reduction in width t-t'

• 로봇학회논문지
• /
• 제3권4호
• /
• pp.270-277
• /
• 2008

The development of a face robot basically targets very natural human-robot interaction (HRI), especially emotional interaction. So does a face robot introduced in this paper, named Buddy. Since Buddy was developed for a mobile service robot, it doesn't have a living-being like face such as human's or animal's, but a typically robot-like face with hard skin, which maybe suitable for mass production. Besides, its structure and mechanism should be simple and its production cost also should be low enough. This paper introduces the mechanisms and functions of mobile face robot named Buddy which can take on natural and precise facial expressions and make dynamic gestures driven by one laptop PC. Buddy also can perform lip-sync, eye-contact, face-tracking for lifelike interaction. By adopting a customized emotional reaction decision model, Buddy can create own personality, emotion and motive using various sensor data input. Based on this model, Buddy can interact probably with users and perform real-time learning using personality factors. The interaction performance of Buddy is successfully demonstrated by experiments and simulations.

• 대한구순구개열학회지
• /
• 제14권1_2호
• /
• pp.45-56
• /
• 2011

Face transplantation has been reported over the last six years, and it started in 2004 with the announcement of Cleveland Clinic granting the world's first IRB approval to proceed with human face transplantation. Composite tissue allografts (CTAs) in the facial region are considered to be more immunogenic than other solid organ transplants, so these kinds of animal experiments were also considered as challengeable activities to the facial reconstructive surgeons. For the better understanding of CTAs in the oral and maxillofacial fields, we reviewed several recent articles about facial composite transplantation animal model, and summarized some knowledges of composite midface allotransplantation model with sensory and motor reinnervation in this review article.

• 한국생활과학회지
• /
• 제14권1호
• /
• pp.143-153
• /
• 2005

The purpose of this research is to study face images according to the difference of facial shape and make-up. A variety of face images can be formulated by computer graphic simulation, combining numerously different facial shapes and make-up styles. In order to check out the diverse images by make-up styles, we applied five forms of eye brows, two types of eye shadows, and three lip shapes to the round-shaped face of a model. The question sheet, used with a operational stimulant in the experiment, contained 28 articles, composed of a pair of bi-ended adjective in 7 point scale. Data were analyzed using Varimax perpendicular rotation method, Duncan's Multiple Range Test, and Three-way ANOVA. After comparing various results of make-up application to various face types, we could find that facial shape, eye-brows, eye-shadow, and lip shapes influence interactively on total facial images. As a result of make-up image perception analyses, a factor structure was divided into mildness, modernness, elegance, and sociableness. Speaking of make-up image in terms of those factors, round form make-up style showed the highest level of mildness. Upward and straight style of make-up had the highest of modernness. Elegance level went highest when eye shadow style was round form and lip style was straight. Lastly, an incurve lip make-up style showed the highest of sociableness.

• Archives of Plastic Surgery
• /
• 제35권3호
• /
• pp.303-308
• /
• 2008

Purpose: The purpose of this study is to develop three-dimensional computerized anthropometry(3DCA) and to compare its reliability and accuracy 3DCA with manual anthropometry(MA) for measurement of lips and nasal deformities in unilateral cleft lips and palate(UCLP) patients. Methods: Samples were consisted of six UCLP patients whose facial plaster models were available immediately before and 3 months after the cleft lip surgery. MA of the facial plaster models was carried out using an electronic caliper. In 3DCA, three-dimensional auto-measuring program was used to digitize landmarks and to measure three-dimensional virtual facial models (3DVFM), which was generated with a laser scanner and 3D virtual modeling program. Intraclass correlation coefficients(ICC) were calculated to evaluate reliability and reproducibility of the variables in both methods, and Wilcoxon's signed rank test was done to investigate the difference in values of the same variables of facial models of each patient between two methods. Results: All ICC values were higher than 0.8, so both methods could be considered reliable. Although most variables showed statistical differences between two methods(p<0.05), differences between mean values were very small and could be considered not significant in clinical situation. Conclusion: In clinical situation, 3DCA can be an objective, reliable and accurate tool for evaluation of lips and nasal deformities in the cleft patients.

• 디지털콘텐츠학회 논문지
• /
• 제7권4호
• /
• pp.235-243
• /
• 2006

본 논문에서는 얼굴의 표정 변화를 인식하여 실시간으로 감정을 표현하는 아바타를 설계하고 구현하는 방법을 제안하였다. 실시간 감정 표현 아바타는 수동으로 아바타의 표정 패턴을 변화를 주는 것이 아니라 웹캠을 이용하여 실시간으로 입 모양 특징을 추출하고 추출한 입 모양의 패턴을 분석한 뒤, 미리 정의된 표정 패턴에 부합되는 것을 찾는다. 그리고 부합된 표정 패턴을 아바타에 적용하여 아바타의 얼굴 표정을 표현한다. 표정 인식 아바타는 웹캠에서 들어온 영상을 모델 접근 방법을 이용하여 보다 빠르게 입 부분을 인식할 수 있도록 하였다. 그리고 표정 패턴 추출은 표정에 따라 입 모양이 변하는 것을 이용하였다. 모델접근 방법을 이용하여 눈의 정보를 찾은 후 그 정보를 이용하여 입 모양을 추출하게 된다. 기본적으로 13가지 입 모양으로 각각의 표정을 유추하고 각 표정 패턴에 맞는 6개의 아바타를 미리 구현하여 보다 빠르게 아바타의 표정을 변할 수 있게 하였다.

• 한국소성가공학회:학술대회논문집
• /
• 한국소성가공학회 2003년도 The 8th Asian Symposium on Precision Forging ASPF
• /
• pp.18-21
• /
• 2003

A behavior of interface of cold forging tool coated by hard film is analyzed by FEM. The indentation of a hemispherical tool of SUJ2 in JIS coated by electroplated coating of chromium into a cylindrical workpiece of S25C in JIS is performed. The proposed constitutive equations are implemented by the non-linear springs model at the interface. Calculated results show that the maximum interfacial normal displacement is observed at an immediately forward from the lip of the indentation, whereas maximum interfacial tangential displacement is at the immediately backward of the lip of the indentation.

• 한국정밀공학회지
• /
• 제18권9호
• /
• pp.140-146
• /
• 2001

Seals in a universal joint bearing are important components reinforcing lubrication performance by holding lubricant and preventing infiltration of dust, moisture, etc.. There is a great difference in seal performance according to seal shape and bonding position. Therefore, in this study, as for both the lip type seal and the O-ring type seal, FE analysis is conducted using Mooned-Rivlin Model. The results show that O-ring type seal does not have any effect of misalignment angle compared with lib type seal, which is more profitable.