• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.2
• /
• pp.127-133
• /
• 2021

A recent study of people physically falling focused on analyzing the motions of the falls using a recurrent neural network (RNN) and a deep learning approach to get good results from detecting 2D human poses from a single color image. In this paper, we investigate a detection method for estimating the position of the head and shoulder keypoints and the acceleration of positional change using the skeletal keypoints information extracted using PoseNet from an image obtained with a low-cost 2D RGB camera, increasing the accuracy of judgments about the falls. In particular, we propose a fall detection method based on the characteristics of post-fall posture in the fall motion-analysis method. A public data set was used to extract human skeletal features, and as a result of an experiment to find a feature extraction method that can achieve high classification accuracy, the proposed method showed a 99.8% success rate in detecting falls more effectively than a conventional, primitive skeletal data-use method.

• Maxillofacial Plastic and Reconstructive Surgery
• /
• v.22 no.3
• /
• pp.351-355
• /
• 2000

Penetrating injuries in the head and neck region are not common but can pose difficult situations to manage properly. In small cross-sectional area, the neck housed many vital structures, such as carotid artery, internal jugular vein, cervical spines, esophagus, laryngotracheal complex and nerves. Because each vital structure is located within the fascial compartments, bleeding into these closed spaces can give rise to compression of surrounding structures, which may result in compromised airway. Therefore, management of the penetrating injuries should be based on the fully understanding of anatomical relationships, accurate clinical examinations, a careful history taking and the proper treatment planning. We present two cases of penetrating injuries in the head and neck region and discuss on the clinical considerations for the proper management with the literature review.

• Journal of Biomedical Engineering Research
• /
• v.39 no.6
• /
• pp.278-283
• /
• 2018

Nowdays, many people suffer from the neck pain due to forward head posture(FHP) and text neck(TN). To assess the severity of the FHP and TN the craniovertebral angle(CVA) is used in clinincs. However, it is difficult to monitor the neck posture using the CVA in daily life. We propose a new method using the cervical flexion angle(CFA) obtained from a wearable sensor to monitor neck posture in daily life. 15 participants were requested to pose FHP and TN. The CFA from the wearable sensor was compared with the CVA observed from a 3D motion camera system to analyze their correlation. The determination coefficients between CFA and CVA were 0.80 in TN and 0.57 in FHP, and 0.69 in TN and FHP. From the monitoring the neck posture while using laptop computer for 20 minutes, this wearable sensor can estimate the CVA with the mean squared error of 2.1 degree.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.06a
• /
• pp.253-255
• /
• 2016

본 논문에서는 다중 지역 이진 패턴(Multi-scale Bock LBP, MB-LBP) 특징과 랜덤 포레스트에 기반한 새로운 기법의 머리 방향 분류 기법을 제안한다. 제안 기법에서는 occlusion 과 조명의 변화에 강인한 분류 정확도를 얻기 위해서 랜덤화된 트리를 학습하는 것을 목표로 한다. 우선, 얼굴 이미지로부터 많은 MB-LBP 특징을 추출하고, 얼굴 영상들을 랜덤하게 입력하고 MB-LBP 크기 파라미터와 같은 랜덤 특징과 블록 좌표들을 사용하여 트리를 생성한다. 게다가 각 노드에서 정보 이득을 최대화 하는 트리의 내부 노드를 생성하기 위해서 uniform LBP 의 특성을 고려한 분할 함수를 개발한다. 랜덤화된 트리는 랜덤 포레스트에 포함되어 있으며 마지막 결정단계에서 Maximum-A-Posteriori criterion 으로 최종 결정을 한다. 실험 결과는 제안 기법이 다양한 조명, 자세, 표현, occlusion 상황에서 기존의 방법보다 개선된 성능으로 머리 방향을 분류 할 수 있음을 보여준다.

• Journal of Korea Multimedia Society
• /
• v.6 no.4
• /
• pp.594-600
• /
• 2003

In this paper, we propose a detection method for facial features in color images with various backgrounds and face poses. To begin with, the proposed method extracts face candidacy region from images with various backgrounds, which have skin-tone color and complex objects, via the color and edge information of face. And then, by using the elliptical shape property of face, we correct a rotation, scale, and tilt of face region caused by various poses of head. Finally, we verify the face using features of face and detect facial features. In our experimental results, it is shown that accuracy of detection is high and the proposed method can be used in pose-invariant face recognition system effectively

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2000.04a
• /
• pp.463-468
• /
• 2000

본 논문에서는 얼굴의 9가지 상태를 인식하고 이를 상태 시퀀스로 생성한 후, 오토마타 기법을 적용하여 13가지(준비, 상측, 하측, 좌측, 우측, 전진, 후퇴, 좌 윙크, 우 윙크, 좌 더블 윙크, 우 더블 윙크, 긍정, 부정) 헤드 제스처를 인식하는 방법을 제안한다. 얼굴 영역을 추출하는 방법에서는 최적의 얼굴색 정보와 적응적 움직임 정보를 이용하여 얼굴 영역을 추출한다. 눈의 후보 영역을 추출하는 방법에서는 소벨 연산자와 투영 기법을 이용한다. 이 때 눈의 후보들을 제거하기 위하여 눈의 기하학적 정보와 눈은 쌍으로 존재한다는 정보를 이용한다. 얼굴의 상태를 인식하기 위해서는 계층적인 특징분석 방법을 사용한다. 13가지 헤드 제스처는 얼굴 상태 인식의 처리에서 생성된 상태 시퀀스를 이용한 오토마타 기법에 의해 인식된다. 실험 결과, 93.3%의 헤드제스처 인식률을 얻었다.

• Proceedings of the IEEK Conference
• /
• 2009.05a
• /
• pp.270-272
• /
• 2009

본 연구에서는 자동차 사고나 뇌졸중 등에 의해 경추 이하의 마비나 손, 발 등의 움직임은 자유롭지 않은 사람들의 컴퓨터 사용을 돕고자 손이나 발을 이용하지 않고 머리의 움직임과 눈의 깜박임만으로 컴퓨터 마우스 제어가 가능한 장치를 제안하였다. 마우스의 좌우, 상하 이동은 각속도 센서를 이용하여 머리의 움직임으로 유발하고, 클릭과 더블 클릭은 광 센서를 시야를 방해하지 않는 위치에 장착하여 감지할 수 있도록 하였다. 제안한 마우스를 Quick Glance를 이용한 것과 비교해 보고 문자를 입력함에 있어 dasher을 이용하는 것과 윈도우에서 제공하는 화상키보드를 이용하는 것을 비교해 보았다. 공간 이동 능력과 이벤트 검출을 비교한 실험에서는 좌우, 상하 이동은 기존 마우스와 비교하여 속도 면에서는 큰 차이는 없었으나, 정확도가 조금 떨어지는 이유로 인하여 소요시간이 $3{\sim}4$배 정도 더 필요하였다. 그러나 Quick glance와 비교결과에서는 약 14%정도 빠랐고 dasher을 이용하여 문자를 입력함에 있어서도 약 32% 이상 빠르게 문자를 입력할 수 있었다. 실험 결과 눈이 쉽게 피로해지는 Quick glance를 이용하는 것보다 제안한 마우스를 사용하는 것이 더 효과적이었다.

• Journal of International Society for Simulation Surgery
• /
• v.2 no.1
• /
• pp.7-9
• /
• 2015

Purpose In the many face-related application such as head pose estimation, 3D face modeling, facial appearance manipulation, the robust and fast facial feature extraction is necessary. We present the facial feature extraction method based on shape regression and feature selection for real-time facial feature extraction. Materials and Methods The facial features are initialized by statistical shape model and then the shape of facial features are deformed iteratively according to the texture pattern which is selected on the feature pool. Results We obtain fast and robust facial feature extraction result with error less than 4% and processing time less than 12 ms. The alignment error is measured by average of ratio of pixel difference to inter-ocular distance. Conclusion The accuracy and processing time of the method is enough to apply facial feature based application and can be used on the face beautification or 3D face modeling.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.46 no.4
• /
• pp.70-76
• /
• 2009

We proposed the device to control a computer mouse with only head movements and eye blinks so that disabilities by car or other accidents can use a computer. The mouse position were estimated from a gyro-sensor which can measure head movements, and the mouse events such as click/double click were from opto sensors which can detect the eyes flicker, respectively. The sensor was mounted on the goggle in order not to disturb the visual field. There was no difference in movement speed between ours and a general mouse, but it required 3$\sim$4 more times in the result of the experiment to evaluate spatial movements and events detection of the proposed mouse because of the low accuracy. We could eliminate cumbersome work to periodically remove the accumulated error and intuitively control the mouse using non-linear relative point method with dead zones. Optical sensors are used in the event detection circuitry designed to remove the influence of the ambient light changes, therefore it was not affected in the change of external light source.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.11
• /
• pp.465-472
• /
• 2022

In this paper, the style synthesis network is trained to generate style-synthesized video through the style synthesis through training Stylegan and the video synthesis network for video synthesis. In order to improve the point that the gaze or expression does not transfer stably, 3D face restoration technology is applied to control important features such as the pose, gaze, and expression of the head using 3D face information. In addition, by training the discriminators for the dynamics, mouth shape, image, and gaze of the Head2head network, it is possible to create a stable style synthesis video that maintains more probabilities and consistency. Using the FaceForensic dataset and the MetFace dataset, it was confirmed that the performance was increased by converting one video into another video while maintaining the consistent movement of the target face, and generating natural data through video synthesis using 3D face information from the source video's face.