• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.239-240
• /
• 2020

마스크 쓴 얼굴에 대해 랜드마크 분석을 진행하기 위해서는 대량의 마스크가 착용된 얼굴 데이터셋이 필요하다. 본 논문에서는 공개된 얼굴 데이터셋에 자동으로 마스크를 합성하여 대량의 마스크를 착용한 얼굴 데이터셋을 생성하는 시스템을 제안한다. 마스크는 얼굴의 많은 부분을 가리는 물체이다. 따라서 마스크를 쓴 얼굴에 대해서는 일반적인 얼굴 데이터셋으로 학습된 landmark detector가 잘 작동하지 않는다. landmark detector가 잘 작동하게 하려면 마스크를 쓴 얼굴에 대해서 학습을 시켜야 한다. 그러나 현재 마스크를 쓴 얼굴 이미지와 풍부한 landmark 정보를 함께 가지고 있는 데이터셋이 존재하지 않기 때문에 학습에 어려움이 있다. 이 문제를 해결하기 위해 마스크 얼굴 이미지 데이터셋을 만들어내는 방법을 제안하고 마스크를 착용한 얼굴에도 잘 작동하는 랜드마크 검출기를 학습시켜 그 효용을 입증하였다.

• Journal of KIISE:Software and Applications
• /
• v.31 no.2
• /
• pp.182-188
• /
• 2004

This paper proposes to synthesize facial images from a few parameters for the pose and the expression of their constituent components. This parameterization makes the representation, storage, and transmission of face images effective. But it is difficult to parameterize facial images because variations of face images show a complicated nonlinear manifold in high-dimensional data space. To tackle this problem, we use an LLE (Locally Linear Embedding) technique for a good representation of face images, where the relationship among face images is preserving well and the projected manifold into the reduced feature space becomes smoother and more continuous. Next, we apply a snake model to estimate face feature values in the reduced feature space that corresponds to a specific pose and/or expression parameter. Finally, a synthetic face image is obtained from an interpolation of several neighboring face images in the vicinity of the estimated feature value. Experimental results show that the proposed method shows a negligible overlapping effect and creates an accurate and consistent synthetic face images with respect to changes of pose and/or expression parameters.

• Journal of KIISE:Software and Applications
• /
• v.31 no.6
• /
• pp.817-828
• /
• 2004

In this paper, we propose vector-based face generation system that uses montage and shading method and preserves designer(artist)'s style. Proposed system generates character's face similar to human face automatically using facial features that extracted from a photograph. In addition, unlike previous face generation system that uses contours, we propose the system is based on color and composes face from facial features and shade extracted from a photograph. Thus, it has advantages that can make more realistic face similar to human face. Since this system is vector-based, the generated character's face has no size limit and constraint. Therefore it is available to transform the shape freely and to apply various facial expressions to 2D face. Moreover, it has distinctiveness with another approaches in point that can keep artist's impression just as it is in result.

• Journal of the Korea Convergence Society
• /
• v.9 no.4
• /
• pp.47-55
• /
• 2018

In this paper, we propose a novel face recognition(FR) method that takes advantage of combining weighted deep local features extracted from multiple Deep Convolutional Neural Networks(DCNNs) learned with a set of facial local regions. In the proposed method, the so-called weighed deep local features are generated from multiple DCNNs each trained with a particular face local region and the corresponding weight represents the importance of local region in terms of improving FR performance. Our weighted deep local features are applied to Joint Bayesian metric learning in conjunction with Nearest Neighbor(NN) Classifier for the purpose of FR. Systematic and comparative experiments show that our proposed method is robust to variations in pose, illumination, and expression. Also, experimental results demonstrate that our method is feasible for improving face recognition performance.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.06a
• /
• pp.73-74
• /
• 2014

최근 들어 얼굴 교체와 같은 영상 합성 기술들이 많은 관심을 받고 있다. 일반적으로 영상을 합성할 때, 영상간 뚜렷한 명암 차이로 인해 부자연스러운 경계가 발생하는데 이를 자연스럽게 제거하는 블렌딩 기술이 필요하다. 본 논문에서는 이러한 문제를 해결하기 위해 적응적 가중치 기반의 영상 블렌딩 방법을 제안한다. 실험 결과, 본 논문에서 제안하는 방법이 얼굴 합성시 발생하는 뚜렷한 경계 부분을 자연스럽게 제거하고, 합성하고자 하는 얼굴의 눈, 코, 입과 같은 주요 부위를 잘 보존하는 것을 확인할 수 있었다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.17 no.7
• /
• pp.767-778
• /
• 1992

Is discussed. A method to extract feature points and to synthesize human facial images In 3-Dmodel-based ceding system, faciai feature points are extracted automatically using some image processing techniques and the known knowledge for human face. A wire frame model matched to human face Is transformed according to the motion of point using the extracted feature points. The synthesized Image Is produced by mapping the texture of initial front view Image onto the trarnsformed wire frame. Experinent results show that the synthesitzed image appears with little unnaturalness.

• Journal of the Institute of Convergence Signal Processing
• /
• v.2 no.4
• /
• pp.16-21
• /
• 2001

In this paper a virtual form of 3 dimensional face is synthesized from the two sheets of 2 dimensional photographs In this case two sheets of 2D face photographs, the front and the side photographs are used First of all a standard model for a general face is created and from this model the feature points which represents a construction of face are densely defined on part of ears. eyes, a nose and a lip but the other parts. for example, forehead, chin and hair are roughly determined because of flat region or the less individual points. Thereafter the side photograph is connected symmetrically on the left and right sides of the front image and it is gradually synthesized by use of affine transformation method. In order to remove the difference of color and brightness from the junction part, a linear interpolation method is used. As a result it is confirmed that the proposed model which general model of a face can be obtain the 3D virtual image of the individual face.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.1256-1257
• /
• 2022

분할 맵 기반 얼굴 합성 분야의 기존 연구들은 피부 영역을 하나의 라벨로 취급한다. 이는 피부 내 미세한 요소를 표현하지 못하며 고해상도 영상 합성 성능 부족이라는 결과를 초래한다. 본 논문에서는 이러한 문제점을 극복하고자 분할 맵에 주름, 모공, 홍조와 같은 피부 요소를 추가하여 이로부터 얼굴 영상을 합성하는 기법을 제안한다. 기존 분할 맵 기반 영상 합성 연구에 제안하는 기법을 적용하여 정량적 및 정성적 비교를 통해 성능이 개선됨을 보인다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06a
• /
• pp.121-123
• /
• 2006

얼굴의 표정은 얼굴의 구성요소 같은 기하학적 정보와 조명이나 주름 같은 세부적인 정보들로 표현된다. 얼굴 표정은 기하학적 변형만으로는 실감적인 표정을 생성하기 힘들기 때문에 기하학적 변형과 더불어 텍스쳐 같은 세부적인 정보도 함께 변형해야만 실감적인 표현을 할 수 있다. 표정비율이미지 (Expression Ratio Image)같은 얼굴 텍스처의 세부적인 정보를 변형하기 위한 기존 방법들은 조명에 따른 피부색의 변화를 정확히 표현할 수 없는 단점이 있다. 따라서 본 논문에서는 이러한 문제를 해결하기 위해 서로 다른 조명 조건에서도 실감적인 표정 텍스처 정보를 적용할 수 있는 비선형 피부색 모델 기반의 표정 합성 방법을 제안한다. 제안된 방법은 동적 외양 모델을 이용한 자동적인 얼굴 특징 추출과 와핑을 통한 표정 변형 단계, 비선형 피부색 변화 모델을 이용한 표정 생성 단계, Euclidean Distance Transform (EDT)에 의해 계산된 혼합 비율을 사용한 원본 얼굴 영상과 생성된 표정의 합성 등 총 3 단계로 구성된다. 실험결과는 제안된 방법이 다양한 조명조건에서도 자연스럽고 실감적인 표정을 표현한다는 것을 보인다.

• Korean Journal of Cognitive Science
• /
• v.13 no.3
• /
• pp.11-21
• /
• 2002

Parke and Waters' model［1］ of muscle-based face deformation was used to develop a system that can synthesize facial expressions when the pleasure-displeasure and arousal-sleep coordinate values of internal states are specified. Facial expressions sampled from a database developed by Chung, Oh, Lee and Byun ［2］ and its underlying model of internal states were used to find rules for face deformation. The internal - state model included dimensional and categorical values of the sampled facial expressions. To find out deformation rules for each of the expressions, changes in the lengths of 21 facial muscles were measured. Then, a set of multiple regression analyses was performed to find out the relationship between the muscle lengths and internal states. The deformation rules obtained from the process turned out to produce natural-looking expressions when the internal states were specified by the pleasure-displeasure and arousal-sleep coordinate values. Such a result implies that the rules derived from a large scale database and regression analyses capturing the variations of individual muscles can be served as a useful and powerful tool for synthesizing facial expressions.