• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.630-633
• /
• 2020

무의식적인 손-얼굴의 접촉으로 인한 감염의 문제점을 해결하기 위해, 얼굴 만지기 행동을 인식할 필요가 있다. 본 연구는 최근 각광을 받는 딥러닝 기술을 이용하여 비디오 영상에서 얼굴 만지기 행동 인식에 대한 연구이다. 우선, 비디오 영상에서 얼굴 만지기와 관련된 11 가지 행동에 대한 시, 공간적 특징을 컨볼루션 신경망을 통해 추출한다. 추출된 정보는 각 행동 레이블로 인코딩되어 비디오 영상에서 얼굴 만지기 행동을 분류한다. 또한, 3D, 2D 컨볼루션 신경망의 대표 네트워크인 I3D, MobileNet v3에 대해 비교 실험을 진행한다. 제안하는 시스템을 적용하여 인간의 행동을 분류하는 실험을 진행했을 때, 얼굴을 만지는 행동을 99%의 확률로 구분했다. 이 시스템을 이용하여 일반인이 무의식적인 얼굴 만지기 행동에 대해서 정량적으로 또는 적시적으로 인식을 하여, 안전한 위생 습관을 확립하여 감염의 확산방지에 도움을 줄수 있기를 바란다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2003.11a
• /
• pp.232-235
• /
• 2003

본 눈문에서는 표면의 에지를 잘 나타내는 웨이블릿을 이용한 3차원 얼굴 인식 알고리듬을 제안한다. 먼저 얼굴영역을 추출하고 정규화과정을 수행한다. 코는 얼굴에서 가장 높고 기준점의 역할을 하므로 반복 선택방법을 이용해서 코끝을 찾는다. 코끝 최고점을 기준으로 깊이값 20, 30, 40인 영역에 대해 웨이블릿 변환을 수행하여 얼굴마다 저주파와 고주파들을 생성하는데, 저주파를 제외한 고주파들에 대하여 히스토그램을 특징벡터로 사용하였다 유사도의 비교는 L$_1$거리함수를 사용하여 수평, 수직, 대각고주파, 그리고 이 고주파들의 유사도 비교치를 합한 합성의 경우 각각에 대하여 실험하였다. 깊이값에 따른 영역에서 고주파별로 실험한 결과, 순위 임계값이 10위를 기준으로 깊이값 30 대각고주파에서 91%가 나타났고 합성에서는 93%의 인식률이 나왔다.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.17 no.2
• /
• pp.141-146
• /
• 2017

New infectious diseases such as MERS have been in need of many measures such as initial discovery, isolation, and crisis response. In addition, the culture of hospitals is changing, such as the general public 's visiting and Nursing Care Integration Services. However, as the qualifications and regulations of medical personnel in hospitals become rigid, overseas such as linens, wastes movements are replacing possible works with robots. we have developed a hospital logistics robot that can carry out various goods delivery within a hospital, and can move various kinds of objects safely to a desired location. In this thesis, we have studied a hospital logistics robot that can carry out various kinds of goods delivery within the hospital, and can move various kinds of objects such as waste, and linen safely to a desired location. The movement of a robot in a hospital may cause a collision between a person and an object, so that the collision must be prevented. In order to prevent collision, it is necessary to recognize whether or not an object exists in the movement path of the robot. And if there is an object, it should recognize whether it moves or not. In order to recognize human beings and objects, we recognize the person with face/body recognition technology and generate the context awareness of the object using 3D Vision image segmentation technology. We use the generated information to create a map that considers objects and person in the robot moving range. Thus, the robot can be operated safely and efficiently.

• Journal of Society of Preventive Korean Medicine
• /
• v.19 no.3
• /
• pp.1-10
• /
• 2015

Objectives : The purpose of this study was to reveal the prevalence of Mibyeong and its symptoms including fatigue, pain, sleep disturbance, dyspepsia, depression, anxiety and anger by using the national survey. Methods : Questionnaires were collected by Gallup Korea. Participants were chosen through stratified sampling method based on area, gender and age. Questionnaire was designated to confirm the recognition, managing of Mibyeong, investigation of life habit, medical history, basic information, QoL questionnaires (Short Form-12, EuroQol-5D) and understanding of Mibyeong medical service conditions. Generally all questionnaires were used for survey the Mibyeong status in public except QoL questionnaires. Questionnaires were fulfilled by professional surveyor as face to face interview. Descriptives was used for data analysis and the results were expressed as percentage ratios (%) Results : 1,101 of people were acquired in this study. Eighty point two (80.2%) percent of participants did not know the concept of Mibyeong accurately even though 80.6% complained of Mibyeong related symptoms. Among them, fatigue was accounted for the highest response (70.7%)in this study. Sixty point four percent of participants identified non-smoking, stop drinking, eating habits and sleeping habits as a way to manage their Mibyeong related symptoms. In addition, exercising (60.8%), visiting medical institution (58.4%) and taking health functional food (52.7%) were presented. Only 23.1% among people with symptoms Mibyeong visited medical facilities. Moreover, the quality of life was found to be significantly correlated with health status. Conclusions : This study could contribute to express the importance of announcing the concept of Mibyeong and status to Korean public. Moreover, more Mibyeong studies should be conducted in the future to evaluate the Mibyeong status objectively.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1997.10a
• /
• pp.41-44
• /
• 1997

This paper is to provide an integration approach between design and process planning for mechanical parts, using feature recognition. We develop a method to extract each individual feature of an object from 3D modeling data using face-edge graph based algorithm and then propose an approach to recognize the volumic form features using heuristic rules. we demonstrate the proposed approaches are effective for such basic shapes as pocket, slot, through hole, etc.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.4
• /
• pp.379-384
• /
• 2009

A human gaze detection and tracing method is importantly required for HMI(Human-Machine-Interface) like a Human-Serving robot. This paper proposed a novel three-dimension (3D) human gaze estimation method by using a face recognition, an orientation estimation and SVMs (Support Vector Machines). 2,400 images with the pan orientation range of $-90^{\circ}{\sim}90^{\circ}$ and tilt range of $-40^{\circ}{\sim}70^{\circ}$ with intervals unit of $10^{\circ}$ were used. A stereo camera was used to obtain the global coordinate of the center point between eyes and Gabor filter banks of horizontal and vertical orientation with 4 scales were used to extract the facial features. The experiment result shows that the error rate of proposed method is much improved than Liddell's.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.640-643
• /
• 2020

사진, 동영상 또는 3D 프린터로 변조된 얼굴을 이용한 안면인식 보안 우회 사례가 증가하였다. 따라서 본 논문에서는 얼굴인증 단계인 본인 여부와 Liveness Detection 및 OTP를 통한 다중인증 프레임워크를 구현함으로써 기존 단일인증 대비 더욱 안전한 인증 환경을 기대한다.

• Journal of Broadcast Engineering
• /
• v.23 no.3
• /
• pp.351-360
• /
• 2018

Human emotion recognition is a research topic that is receiving continuous attention in computer vision and artificial intelligence domains. This paper proposes a method for classifying human emotions through multiple neural networks based on multi-modal signals which consist of image, landmark, and audio in a wild environment. The proposed method has the following features. First, the learning performance of the image-based network is greatly improved by employing both multi-task learning and semi-supervised learning using the spatio-temporal characteristic of videos. Second, a model for converting 1-dimensional (1D) landmark information of face into two-dimensional (2D) images, is newly proposed, and a CNN-LSTM network based on the model is proposed for better emotion recognition. Third, based on an observation that audio signals are often very effective for specific emotions, we propose an audio deep learning mechanism robust to the specific emotions. Finally, so-called emotion adaptive fusion is applied to enable synergy of multiple networks. The proposed network improves emotion classification performance by appropriately integrating existing supervised learning and semi-supervised learning networks. In the fifth attempt on the given test set in the EmotiW2017 challenge, the proposed method achieved a classification accuracy of 57.12%.

• Journal of the Society of Cosmetic Scientists of Korea
• /
• v.39 no.4
• /
• pp.323-327
• /
• 2013

Various indicators representing skin characteristics such as skin hydration, sebum excretion rate, lightness, and pH are different depending on environmental and genetic factors. However, they are absolute skin indicators and are different from skin characteristics that a person recognizes. Based on this fact, many recent studies have been mainly conducting researches on perspective changes according to changes of absolute skin. This study was proposed not only to find out differences on skin colors of asian by nations, but also to find out whether there was any difference in skin brightness they perceive depending on actual skin color changing. As many as 410 subjects of three Asia nations were participated in this study, and investigated their responses on skin brightness using questionnaire, which was answered their skin color in three different levels. It was also were analyzed how actual skin brightness were changed depending on their perceived skin color changes of subjects. There was a trend showing that the brightness of the actual skin color was increased when participants felt their skin color got brighter regardless of their nationalities. However, there were some differences in color between perceived color and actual color. In addition, there was a different aspect by nations in changes of skin redness and skin yellowness. In conclusion, it was revealed that factors which help people to perceive their own skin brightness were not based on absolute skin brightness, but on different criteria depending on where they are from.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.3
• /
• pp.247-253
• /
• 2019

In this paper, we present a new method for real-time tracking of human walking around a laser sensor system. The method converts range data with $r-{\theta}$ coordinates to a 2D image with x-y coordinates. Then human tracking is performed using human's features, i.e. appearances of human walking pattern, and the input range data. The laser sensor based human tracking method has the advantage of simplicity over conventional methods which extract human face in the vision data. In our method, the problem of estimating 2D positions and orientations of two walking human's ankle level is formulated based on a moving trajectory algorithm. In addition, the proposed tracking system employs a HMM to robustly track human in case of occlusions. Experimental results using a real system demonstrate usefulness of the proposed method.