• Journal of Information Processing Systems
• /
• v.17 no.2
• /
• pp.399-410
• /
• 2021

Aiming at the problem that complex external variables in natural scenes have a greater impact on facial expression recognition results, a facial expression recognition method based on two-stream convolutional neural network is proposed. The model introduces exponentially enhanced shared input weights before each level of convolution input, and uses soft attention mechanism modules on the space-time features of the combination of static and dynamic streams. This enables the network to autonomously find areas that are more relevant to the expression category and pay more attention to these areas. Through these means, the information of irrelevant interference areas is suppressed. In order to solve the problem of poor local robustness caused by lighting and expression changes, this paper also performs lighting preprocessing with the lighting preprocessing chain algorithm to eliminate most of the lighting effects. Experimental results on AFEW6.0 and Multi-PIE datasets show that the recognition rates of this method are 95.05% and 61.40%, respectively, which are better than other comparison methods.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.1121-1122
• /
• 2008

By using FACS(Facial Action Coding System) and linear interpolation, a 3D facial robot simulator is developed in this paper. This simulator is based on real facial robot and synchronizes with it by unifying protocol. Using AUs(Action Unit) of each 5 basic expressions and linear interpolation makes more various dynamic facial expressions.

• Journal of Institute of Control, Robotics and Systems
• /
• v.16 no.10
• /
• pp.957-962
• /
• 2010

If we want to recognize the human's emotion via the facial image, first of all, we need to extract the emotional features from the facial image by using a feature extraction algorithm. And we need to classify the emotional status by using pattern classification method. The AAM (Active Appearance Model) is a well-known method that can represent a non-rigid object, such as face, facial expression. The Bayesian Network is a probability based classifier that can represent the probabilistic relationships between a set of facial features. In this paper, our approach to facial feature extraction lies in the proposed feature extraction method based on combining AAM with FACS (Facial Action Coding System) for automatically modeling and extracting the facial emotional features. To recognize the facial emotion, we use the DBNs (Dynamic Bayesian Networks) for modeling and understanding the temporal phases of facial expressions in image sequences. The result of emotion recognition can be used to rehabilitate based on biofeedback for emotional disabled.

• Journal of the Korea Computer Graphics Society
• /
• v.18 no.3
• /
• pp.9-16
• /
• 2012

In this paper, we present a new method for reconstructing detailed facial expression from roughly captured data with a small number of markers. Because of the difference in the required capture resolution between the full-body capture and the facial expression capture, they hardly have been performed simultaneously. However, for generating natural animation, a simultaneous capture for body and face is essential. For this purpose, we provide a method for capturing the detailed facial expression only with a small number of markers. Our basic idea is to build a database for the facial expressions and apply the principal component analysis for reducing the dimensionality. The dimensionality reduction enables us to estimate the full data from a part of the data. We justify our method by applying it to dynamic scenes to show the viability of the method.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.1
• /
• pp.1-5
• /
• 2009

This paper suggests to apply mirror-reflected method based 2D emotion recognition database to 3D application. Also, it makes facial expression of fuzzy modeling using probability of emotion. Suggested facial expression function applies fuzzy theory to 3 basic movement for facial expressions. This method applies 3D application to feature vector for emotion recognition from 2D application using mirror-reflected multi-image. Thus, we can have model based on fuzzy nonlinear facial expression of a 2D model for a real model. We use average values about probability of 6 basic expressions such as happy, sad, disgust, angry, surprise and fear. Furthermore, dynimic facial expressions are made via fuzzy modelling. This paper compares and analyzes feature vectors of real model with 3D human-like avatar.

• Proceedings of the IEEK Conference
• /
• 2002.07c
• /
• pp.1760-1763
• /
• 2002

In this paper the authors present their attempt io realize a motion pattern detector that finds specified sequence of image from input motion image. The detector is intended to be used for time-varying facial expression understanding. Needless to say, facial expression understanding by machine is crucial and enriches quality of human machine interaction. Among various facial expressions, like blinking, there must be such expressions that can not be recognized if input expression image is static. Still image of blinking can not be distinguished from sleeping. In this paper, the authors discuss implementation of their motion pattern detector and describe experiments using the detector. Experimental results confirm the feasibility of the idea behind the implemented detector.

• Korean Journal of Cognitive Science
• /
• v.14 no.1
• /
• pp.10-10
• /
• 2003

This paper extracts the edge of main components of face with Gabor wavelets transformation in facial expression images. FCM(Fuzzy C-Means) clustering algorithm then extracts the representative feature points of low dimensionality from the edge extracted in neutral face. The feature-points of the neutral face is used as a template to extract the feature-points of facial expression images. To match point to Point feature points on an expression face against each feature point on a neutral face, it consists of two steps using a dynamic linking model, which are called the coarse mapping and the fine mapping. This paper presents an automatic extraction of feature-points by dynamic linking model based on Gabor wavelets and fuzzy C-means(FCM) algorithm. The result of this study was applied to extract features automatically in facial expression recognition based on dimension[1].

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.04a
• /
• pp.401-404
• /
• 2007

본 논문은 거울 투영을 이용하여 2D의 감정인식 데이터베이스를 3D에 적용 가능하다는 것을 증명한다. 또한, 감정 확률을 이용하여 퍼지 모델링을 기반으로한 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 6가지인 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.

• The KIPS Transactions:PartA
• /
• v.11A no.5
• /
• pp.383-390
• /
• 2004

This paper propose a method that controls facial expression of 3D avatar by having the user select a sequence of facial expressions in the space of facial expressions. And we setup its system. The space of expression is created from about 2400 frames consist of motion captured data of facial expressions. To represent the state of each expression, we use the distance matrix that represents the distances between pairs of feature points on the face. The set of distance matrices is used as the space of expressions. But this space is not such a space where one state can go to another state via the straight trajectory between them. We derive trajectories between two states from the captured set of expressions in an approximate manner. First, two states are regarded adjacent if the distance between their distance matrices is below a given threshold. Any two states are considered to have a trajectory between them If there is a sequence of adjacent states between them. It is assumed . that one states goes to another state via the shortest trajectory between them. The shortest trajectories are found by dynamic programming. The space of facial expressions, as the set of distance matrices, is multidimensional. Facial expression of 3D avatar Is controled in real time as the user navigates the space. To help this process, we visualized the space of expressions in 2D space by using the multidimensional scaling(MDS). To see how effective this system is, we had users control facial expressions of 3D avatar by using the system. As a result of that, users estimate that system is very useful to control facial expression of 3D avatar in real-time.

• The Journal of Korea Robotics Society
• /
• v.13 no.1
• /
• pp.16-25
• /
• 2018

The talking head (TH) indicates an utterance face animation generated based on text and voice input. In this paper, we propose the generation method of TH with facial expression and intonation by speech input only. The problem of generating TH from speech can be regarded as a regression problem from the acoustic feature sequence to the facial code sequence which is a low dimensional vector representation that can efficiently encode and decode a face image. This regression was modeled by bidirectional RNN and trained by using SAVEE database of the front utterance face animation database as training data. The proposed method is able to generate TH with facial expression and intonation TH by using acoustic features such as MFCC, dynamic elements of MFCC, energy, and F0. According to the experiments, the configuration of the BLSTM layer of the first and second layers of bidirectional RNN was able to predict the face code best. For the evaluation, a questionnaire survey was conducted for 62 persons who watched TH animations, generated by the proposed method and the previous method. As a result, 77% of the respondents answered that the proposed method generated TH, which matches well with the speech.