• 한국멀티미디어학회논문지
• /
• 제17권9호
• /
• pp.1076-1083
• /
• 2014

We propose an emotional facial avatar that portrays the user's facial expressions with an emotional emphasis, while achieving visual and behavioral realism. This is achieved by unifying automatic analysis of facial expressions and animation of realistic 3D faces with details such as facial hair and hairstyles. To augment facial appearance according to the user's emotions, we use emotional templates representing typical emotions in an artistic way, which can be easily combined with the skin texture of the 3D face at runtime. Hence, our interface gives the user vision-based control over facial animation of the emotional avatar, easily changing its moods.

• 한국지능시스템학회논문지
• /
• 제19권6호
• /
• pp.821-827
• /
• 2009

표정은 인간의 감정을 전달할 수 있는 중요한 수단으로 표정 인식은 감정상태를 알아낼 수 있는 효과적인 방법중 하나이다. 일반적인 표정 인식 시스템은 얼굴 표정을 표현하는 특징점을 찾고, 물리적인 해석 없이 특징을 추출한다. 하지만 특징점 추출은 많은 시간이 소요될 뿐 아니라 특징점의 정확한 위치를 추정하기 어렵다. 그리고 표정 인식 시스템을 실시간 임베디드 시스템에서 구현하기 위해서는 알고리즘을 간략화하고 자원 사용량을 줄일 필요가 있다. 본 논문에서 제안하는 실시간 표정 인식 시스템은 격자점 위치에서 얻어진 가버 웨이블릿(Gabor wavelet) 특징 기반 표정 공간을 설정하고, 각 표정 공간에서 얻어진 주성분을 신경망 분류기를 이용하여 얼굴 표정을 분류한다. 제안하는 실시간 표정 인식 시스템은 화남, 행복, 평온, 슬픔 그리고 놀람의 5가지 표정이 인식 가능하며, 다양한 실험에서 평균 10.25ms의 수행시간, 그리고 87%~93%의 인식 성능을 보였다.

• 한국지능시스템학회논문지
• /
• 제19권1호
• /
• pp.1-5
• /
• 2009

본 논문은 거울 투영을 이용하여 2D의 감정인식 데이터베이스를 3D에 적용 가능하다는 것을 증명한다. 또한, 감정 확률을 이용하여 퍼지 모델링 기반의 얼굴표정을 생성하고, 표정을 움직이는 3가지 기본 움직임에 대한 퍼지이론을 적용하여 얼굴표현함수를 제안한다. 제안된 방법은 거울 투영을 통한 다중 이미지를 이용하여 2D에서 사용되는 감정인식에 대한 특징벡터를 3D에 적용한다. 이로 인해, 2D의 모델링 대상이 되는 실제 모델의 기본감정에 대한 비선형적인 얼굴표정을 퍼지를 기반으로 모델링한다. 그리고 얼굴표정을 표현하는데 기본 감정 5가지인 행복, 슬픔, 혐오, 화남, 놀람, 무서움으로 표현되며 기본 감정의 확률에 대해서 각 감정의 평균값을 사용하고 6가지 감정 확률을 이용하여 동적 얼굴표정을 생성한다. 제안된 방법을 3D 인간형 아바타에 적용하여 실제 모델의 표정 벡터와 비교 분석한다.

• 한국콘텐츠학회논문지
• /
• 제17권2호
• /
• pp.439-448
• /
• 2017

얼굴 표정은 상호간의 소통에 있어 중요한 의미를 갖는다. 얼굴 근육의 움직임은 감성 정보를 제공하는데, 이는 사회적 관계를 향상하는 데 중요한 역할을 한다. 그러나 얼굴의 단순 움직임만으로는 복잡한 사회 감성을 인식하기에 정확하지 않다. 본 연구의 목적은 친밀도, 공감도, 긍정도와 같은 사회감성을 인식하기 위한 얼굴의 미세 움직임을 분석하는 것이다. 76명의 피험자를 대상으로 상기 사회감성을 유발하는 자극을 제시하였고 카메라를 사용하여 얼굴 표정을 측정하였다. 결론적으로 친밀함, 공감도, 긍정도의 사회 감성에서 얼굴의 미세움직임이 다르게 나타났다. 총 44개의 얼굴 근육 중 3개의 무의식 근육과 18개의 의식 근육의 움직임 양을 추출한 후, 고속푸리에변환(Fast Fourier Tranform, FFT)을 통하여 (Dominant) Frequency 대역을 확인하였다. 독립 t-검정 결과, 친밀도 상황에서는 코 주변과 볼 주변 근육, 공감도 상황에서는 입 주변 근육, 긍정도 상황에서는 턱 주변 근육에서 유의한 차이를 보였다. 이는 애니메이션의 가상 아바타 등 얼굴 표정의 새로운 표현요소를 제안하고 근육에 따른 사회감성 인식의 기초 연구로서 활용 가능할 것으로 사료 된다.

• 융합보안논문지
• /
• 제14권7호
• /
• pp.23-28
• /
• 2014

본 논문에서는 기존에 제안된 LDP(Local Directional Pattern)를 기반으로 얼굴 표정 인식 시스템에 대한 설계 및 구현 방법을 제안한다. LDP는 얼굴 영상을 구성하고 있는 각 화소를 주변 화소들과의 관계를 고려하여 지역적인 미세 패턴(Local Micro Pattern)으로 표현해준다. 새롭게 제시된 LDP에서 생성되는 코드들이 다양한 조건하에서 정확한 정보를 포함할 수 있는지의 여부를 검증할 필요가 있다. 따라서, 새롭게 제안된 지역 미세 패턴인 LDP를 다양한 환경에서 신속하게 검증하기 위한 평가 시스템을 구축한다. 제안된 얼굴 표정인식 평가 시스템에서는 6개의 컴포넌트를 거쳐 얼굴 표정인식률을 계산할 수 있도록 구성하였으며, Gabor, LBP와 비교하여 LDP의 인식률을 검증한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권3호
• /
• pp.1390-1403
• /
• 2016

Facial expression recognition (FER) plays a very significant role in computer vision, pattern recognition, and image processing applications such as human computer interaction as it provides sufficient information about emotions of people. For video-based facial expression recognition, depth cameras can be better candidates over RGB cameras as a person's face cannot be easily recognized from distance-based depth videos hence depth cameras also resolve some privacy issues that can arise using RGB faces. A good FER system is very much reliant on the extraction of robust features as well as recognition engine. In this work, an efficient novel approach is proposed to recognize some facial expressions from time-sequential depth videos. First of all, efficient Local Binary Pattern (LBP) features are obtained from the time-sequential depth faces that are further classified by Generalized Discriminant Analysis (GDA) to make the features more robust and finally, the LBP-GDA features are fed into Hidden Markov Models (HMMs) to train and recognize different facial expressions successfully. The depth information-based proposed facial expression recognition approach is compared to the conventional approaches such as Principal Component Analysis (PCA), Independent Component Analysis (ICA), and Linear Discriminant Analysis (LDA) where the proposed one outperforms others by obtaining better recognition rates.

• Asian Journal for Public Opinion Research
• /
• 제8권3호
• /
• pp.180-196
• /
• 2020

This paper considers Japanese political interviews to integrate conversation and facial expression analysis. The behaviors of political leaders will be disclosed by analyzing questions and responses by using the turn-taking system in conversation analysis. Additionally, audiences who cannot understand verbal expressions alone will understand the psychology of political leaders by analyzing their facial expressions. Integral analyses promote understanding of the types of facial and verbal expressions of politicians and their effect on public opinion. Politicians have unique techniques to convince people. If people do not know these techniques and ways of various expressions, they will become confused, and politics may fall into populism as a result. To avoid this, a complete understanding of verbal and non-verbal behaviors is needed. This paper presents two analyses. The first analysis is a qualitative analysis that deals with Prime Minister Shinzō Abe and shows that differences between words and happy facial expressions occur. That result indicates that Abe expresses disgusted facial expressions when faced with the same question from an interviewer. The second is a quantitative multiple regression analysis where the dependent variables are six facial expressions: happy, sad, angry, surprised, scared, and disgusted. The independent variable is when politicians have a threat to face. Political interviews that directly inform audiences are used as a tool by politicians. Those interviews play an important role in modelling public opinion. The audience watches political interviews, and these mold support to the party. Watching political interviews contributes to the decision to support the political party when they vote in a coming election.

• 제어로봇시스템학회논문지
• /
• 제13권4호
• /
• pp.371-376
• /
• 2007

This paper presents an emotion recognition and its expression system of an intelligent robot like a home robot or a service robot. Emotion recognition method in the robot is used by a facial image. We use a motion and a position of many facial features. apply a tracking algorithm to recognize a moving user in the mobile robot and eliminate a skin color of a hand and a background without a facial region by using the facial region detecting algorithm in objecting user image. After normalizer operations are the image enlarge or reduction by distance of the detecting facial region and the image revolution transformation by an angel of a face, the mobile robot can object the facial image of a fixing size. And materialize a multi feature selection algorithm to enable robot to recognize an emotion of user. In this paper, used a multi layer perceptron of Artificial Neural Network(ANN) as a pattern recognition art, and a Back Propagation(BP) algorithm as a learning algorithm. Emotion of user that robot recognized is expressed as a graphic LCD. At this time, change two coordinates as the number of times of emotion expressed in ANN, and change a parameter of facial elements(eyes, eyebrows, mouth) as the change of two coordinates. By materializing the system, expressed the complex emotion of human as the avatar of LCD.

• Journal of Information Processing Systems
• /
• 제10권3호
• /
• pp.443-458
• /
• 2014

An extreme learning machine (ELM) is a recently proposed learning algorithm for a single-layer feed forward neural network. In this paper we studied the ensemble of ELM by using a bagging algorithm for facial expression recognition (FER). Facial expression analysis is widely used in the behavior interpretation of emotions, for cognitive science, and social interactions. This paper presents a method for FER based on the histogram of orientation gradient (HOG) features using an ELM ensemble. First, the HOG features were extracted from the face image by dividing it into a number of small cells. A bagging algorithm was then used to construct many different bags of training data and each of them was trained by using separate ELMs. To recognize the expression of the input face image, HOG features were fed to each trained ELM and the results were combined by using a majority voting scheme. The ELM ensemble using bagging improves the generalized capability of the network significantly. The two available datasets (JAFFE and CK+) of facial expressions were used to evaluate the performance of the proposed classification system. Even the performance of individual ELM was smaller and the ELM ensemble using a bagging algorithm improved the recognition performance significantly.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권1호
• /
• pp.196-212
• /
• 2014

Facial expression recognition is an important part in emotional interaction between human and machine. This paper proposes a facial expression recognition approach based on multi-classifier fusion with stacking algorithm. The kappa-error diagram is employed in base-level classifiers selection, which gains insights about which individual classifier has the better recognition performance and how diverse among them to help improve the recognition accuracy rate by fusing the complementary functions. In order to avoid the influence of the chance factor caused by guessing in algorithm evaluation and get more reliable awareness of algorithm performance, kappa and informedness besides accuracy are utilized as measure criteria in the comparison experiments. To verify the effectiveness of our approach, two public databases are used in the experiments. The experiment results show that compared with individual classifier and two other typical ensemble methods, our proposed stacked ensemble system does recognize facial expression more accurately with less standard deviation. It overcomes the individual classifier's bias and achieves more reliable recognition results.