• IEIE Transactions on Smart Processing and Computing
• /
• v.2 no.6
• /
• pp.323-331
• /
• 2013

An extreme learning machine (ELM) is an efficient learning algorithm that is based on the generalized single, hidden-layer feed-forward networks (SLFNs), which perform well in classification applications. Many studies have demonstrated its superiority over the existing classical algorithms: support vector machine (SVM) and BP neural network. This paper presents a novel face recognition approach based on a multi-class project extreme learning machine (MPELM) classifier and 2D Gabor transform. First, all face image features were extracted using 2D Gabor filters, and the MPELM classifier was used to determine the final face classification. Two well-known face databases (CMU-PIE and ORL) were used to evaluate the performance. The experimental results showed that the MPELM-based method outperformed the ELM-based method as well as other methods.

• Proceedings of the IEEK Conference
• /
• 2002.07b
• /
• pp.1252-1255
• /
• 2002

In this paper, face recognition algorithm, by using skin color information of HSI color coordinate collected from face images, elliptical mask, fratures of face including eyes, nose and mouth, and geometrical feature vectors of face and facial angles, is proposed. The proposed algorithm improved face region extraction efficacy by using HSI information relatively similar to human's visual system along with color tone information about skin colors of face, elliptical mask and intensity information. Moreover, it improved face recognition efficacy with using feature information of eyes, nose and mouth, and Θ1(ACRED), Θ2(AMRED) and Θ 3(ANRED), which are geometrical face angles of face. In the proposed algorithm, it enables exact face reading by using color tone information, elliptical mask, brightness information and structural characteristic angle together, not like using only brightness information in existing algorithm. Moreover, it uses structural related value of characteristics and certain vectors together for the recognition method.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2004.05a
• /
• pp.106-109
• /
• 2004

Many reported methods assume that the faces in an image or an image sequence have been identified and localization. Face detection from image is a challenging task because of variability in scale, location, orientation and pose. In this paper, we present an efficient linear discriminant for multi-view face detection. Our approaches are based on linear discriminant. We define training data with fisher linear discriminant to efficient learning method. Face detection is considerably difficult because it will be influenced by poses of human face and changes in illumination. This idea can solve the multi-view and scale face detection problem poses. Quickly and efficiently, which fits for detecting face automatically. In this paper, we extract face using fisher linear discriminant that is hierarchical models invariant pose and background. We estimation the pose in detected face and eye detect. The purpose of this paper is to classify face and non-face and efficient fisher linear discriminant..

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.07a
• /
• pp.29-30
• /
• 2021

본 논문에서는 이미지 분류에 쓰이는 최신 모델로 CNN과 ImageNet을 기반으로 한 EfficientNet을 활용해서 Square, Oval, Oblong, Round, Heart 총 다섯 가지의 얼굴 모양으로 분류하는 task에 두 가지 데이터로 실험해보고 추가적으로 Image Augmentation 기법을 활용해 성능향상을 보였다.

• Journal of Sasang Constitutional Medicine
• /
• v.29 no.2
• /
• pp.95-103
• /
• 2017

Objectives This study was aimed to analyze characteristics of facial shapes in adult women by sasang constitution using hyungsang classification. Methods Using a digital camera, we took a picture of 1,011 women who participated in clinical study on menstrual pain and acquired their 3D facial images with a face-only scanner. They filled out SSCQ-P(sasang constitution questionnaire for patient) for the diagnosis of sasang constitution. Based on the above photographs and 3D images, one of the hyungsang medicine specialist diagnosed according to five diagnostic criteria. The sasang constitution was diagnosed by referring to questionnaires and photographs. Frequency analysis was performed using the statistical analysis system version 9.4 and chi-square test was performed for validity evaluation. Results In taeeumin, the wide face shape(n=261, 74.36%) was much more than the narrow shape(n=90, 25.64%) and the convex face profile(n=164, 85.86%) was much more than the concave profile(n=27, 14.14%). Regardless of sasang constitution, angular face shape(n=501, 50%) was the most, followed by oval shape(n=317, 31.64%). Subjects with big ears(n=291, 29.19%) were the most, while big eyes(n=104, 10.43%) were the least. Subjects with eyes and nose tip upward(n=615, 78.05%) were the most, while eyes and nose tip downward(n=22, 2.79%) were the least. Conclusions Most Korean adult women have angular face, such as square or diamond, with slanted eyes and upturned nose. Taeeumin women have wide facial shape and convex profile.

• Journal of the Korean Society for Library and Information Science
• /
• v.55 no.4
• /
• pp.267-291
• /
• 2021

Due to the global pandemic caused by COVID-19 in 2020, there have been major changes in the education sites. Universities have fully introduced remote learning, which was considered as an auxiliary education, and non-face-to-face classes have become commonplace, and professors and students are making great efforts to adapt to the new educational environment. In order to improve the quality of non-face-to-face lectures amid these changes, it is necessary to study the factors affecting lecture satisfaction. Therefore, This paper presents a new methodology using big data to identify the factors affecting university lecture satisfaction changed before and after COVID-19. We use Topic Modeling method to analyze lecture reviews before and after COVID-19, and identify factors affecting lecture satisfaction. Through this, we suggest the direction for university education to move forward. In addition, we can identify the factors of satisfaction and dissatisfaction of lectures from multiangle by establishing a topic classification model with an F1-score of 0.84 based on KoBERT, a deep learning language model, and further contribute to continuous qualitative improvement of lecture satisfaction.

• The Research Journal of the Costume Culture
• /
• v.8 no.1
• /
• pp.133-144
• /
• 2000

The purpose of this study was to classify women's face types by visual distinction and to analyze the measurement of face types. A survey was conducted by subjects of 167 women's college students in Kwangju City and Chonnam area. Data were analyzed by Frequencies, Mean, one way ANOVA and Ducan's Multiple Range Test. The major results were as followed ; ·Women's face types were classified by 7 types and there were oblong shape(28.3%), egg shape(25.7%), round shape(23.9%), square shape(12.4%), inverted triangle shape(5.3%), diamond shape(3.5%), triangle shape(0.8%) in the subjects. ·From the measurements of the women's face, index of face length to face breadth was 1.38, it means that the index was different from the other refferences. And the lower face length was longer than the upper and the middle face lengths. ·Differences From those measurements like forehead breadth, face length/bizigion breath(p〈.001), bizigion breadth, bignathion slopper, stature(p〈.01) and trichion breadth, tragion-menton length(p〈.05) were significant in the classified face types.

• The Research Journal of the Costume Culture
• /
• v.12 no.6 s.53
• /
• pp.933-944
• /
• 2004

This study was aimed to provide the fundamental and various measurement data of the head and face for male children. Two hundred forty one male children, aged nine to twelve years, participated for this study. The 31 regions on the head and face of the subjects were directly measured by the expert experimenters. Through factor analysis, the six factors were extracted upon factor scores and those factors comprised $67.47\%$ for the total variances. The first factor was described the general height elements for the mouth and the environs of the mouth. The second factor was described the general height around the nose, forehead and eyes. The third factor was described the height of the ear environs. The forth factor contained the length around the sinciput to the occiput, the head thick and the head circumstance. The fifth factor was described the general width of the outer head and the corner of the eyes. The last factor contained the depth of the mouth and nose. Four clusters as their head and face shape were categorized using six factor scores by cluster analysis. Type 1 was characterized by the shortest head and face width, surface length and girth, and the shorter length of head, but the highest position of chin, philtrum, upper lip. Type 2 was characterized by the shortest head and face length and thickness, and the lowest position of the forehead, eye, nose, mouth, ear environs, but that had wider width of head and face. Type 3 was characterized by the longest and the widest head and face type, and the highest position of the mouth. Type 4 was characterized by longer length of head and face, and the widest head girth and largest head thickness, and the highest position of the forehead, eye, nose environs. And this type had the widest width of nose and mouth, and the longest head surface length.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.1
• /
• pp.402-410
• /
• 2017

For many years, researchers in face description area have been representing and recognizing faces based on different methods that include subspace discriminant analysis, statistical learning and non-statistics based approach etc. But still automatic face recognition remains an interesting but challenging problem. This paper presents a novel and efficient face image representation method based on Optimized Local Ternary Pattern (OLTP) texture features. The face image is divided into several regions from which the OLTP texture feature distributions are extracted and concatenated into a feature vector that can act as face descriptor. The recognition is performed using nearest neighbor classification method with Chi-square distance as a similarity measure. Extensive experimental results on Yale B, ORL and AR face databases show that OLTP consistently performs much better than other well recognized texture models for face recognition.

• Journal of Korea Multimedia Society
• /
• v.20 no.8
• /
• pp.1175-1186
• /
• 2017

Human emotion recognition is one of the promising applications in the era of artificial super intelligence. Thus far, facial expression traits are considered to be the most widely used information cues for realizing automated emotion recognition. This paper proposes a novel facial expression recognition (FER) method that works well for recognizing emotion from image sequences. To this end, we develop the so-called weighted soft voting classification (WSVC) algorithm. In the proposed WSVC, a number of classifiers are first constructed using different and multiple feature representations. In next, multiple classifiers are used for generating the recognition result (namely, soft voting) of each face image within a face sequence, yielding multiple soft voting outputs. Finally, these soft voting outputs are combined through using a weighted combination to decide the emotion class (e.g., anger) of a given face sequence. The weights for combination are effectively determined by measuring the quality of each face image, namely "peak expression intensity" and "frontal-pose degree". To test the proposed WSVC, CK+ FER database was used to perform extensive and comparative experimentations. The feasibility of our WSVC algorithm has been successfully demonstrated by comparing recently developed FER algorithms.