• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.68-71
• /
• 2020

최근 딥러닝을 이용한 인체 자세 추정(human pose estimation) 연구가 활발히 진행되고 있다. 그 중 구조가 간단하면서도 성능이 강력하여 널리 사용되고 있는 딥러닝 네트워크 모델은 이미지 분류(image classification)에 사용되는 백본 네트워크(backbone network)와 디컨볼루션 출력망(deconvolution head network)을 이어 붙인 구조를 갖는다[1]. 기존의 디컨볼루션 출력망은 디컨볼루션 층을 쌓아 낮은 해상도의 특징맵을 모두 높은 해상도로 변환한 후 최종 인체 자세 추정을 하는데 이는 다양한 해상도에서 얻어낸 특징들을 골고루 활용하기 힘들다는 단점이 있다. 따라서 본 논문에서는 매 디컨볼루션 층 이후에 인체 자세 추정을 하여 다양한 해상도에서 연산을 하고 이를 종합하여 최종 인체 자세 추정을 하는 방법을 제안한다. 실험 결과 Res50 과 기존의 디컨볼루션 출력망의 경우 0.717 AP 를 얻었는데 Res101 과 기존의 디컨볼루션 출력망을 사용한 결과 50% 이상의 파라미터 수 증가와 함께 0.727 AP, 즉 0.010AP 의 성능 향상이 이루어졌다. 이에 반해 Res50 에 다중 해상도 디컨볼루션 출력망을 사용한 결과 약 1%의 파라미터 수 증가 만으로 0.720 AP, 즉 0.003 AP 의 성능 향상이 이루어졌다. 이를 통해 디컨볼루션 출력망 구조를 개선하면 매우 적은 파라미터 수 증가 만으로도 인체 자세 추정의 성능을 효과적으로 향상시킬 수 있음을 확인하였다.

• Proceedings of the Korean Society for Technology of Plasticity Conference
• /
• 2008.05a
• /
• pp.345-346
• /
• 2008

Most pedestrian-vehicle crashes involve frontal impacts, and the vehicle front structures are responsible for most pedestrian injuries. The vehicle bumper contacts the lower legs at first. The leading edge of the hood (bonnet) strikes the proximal upper leg and finally, the head and upper torso hit the top surface of the hood or windscreen. In essence, the pedestrian wraps around the front of the vehicle until pedestrian and vehicle are traveling at the same speed. Since the hood surface is made from sheet metal, it is a relatively compliant structure and does not pose a major risk for severe head trauma. However, serious head injury can occur when the head hits a region of the hood with stiff underlying structures such as engine components. The solution is to provide sufficient clearance between the hood and underlying structures for controlled deceleration of a pedestrian's head. However, considerations of aerodynamic design and styling can make it extremely difficult to alter a vehicle's front end geometry to provide more under-hood space. In this study, the safe hood will be developed by designing new conceptual inner panel in order to decrease the pedestrian's head injuries without changing hood outer geometry.

• Journal of Internet Computing and Services
• /
• v.21 no.3
• /
• pp.113-121
• /
• 2020

Human Pose Estimation (HPE) which localizes the human body joints becomes a high potential for high-level applications in the field of computer vision. The main challenges of HPE in real-time are occlusion, illumination change and diversity of pose appearance. The single RGB image is fed into HPE framework in order to reduce the computation cost by using depth-independent device such as a common camera, webcam, or phone cam. However, HPE based on the single RGB is not able to solve the above challenges due to inherent characteristics of color or texture. On the other hand, depth information which is fed into HPE framework and detects the human body parts in 3D coordinates can be usefully used to solve the above challenges. However, the depth information-based HPE requires the depth-dependent device which has space constraint and is cost consuming. Especially, the result of depth information-based HPE is less reliable due to the requirement of pose initialization and less stabilization of frame tracking. Therefore, this paper proposes a new method of HPE which is robust in estimating self-occlusion. There are many human parts which can be occluded by other body parts. However, this paper focuses only on head self-occlusion. The new method is a combination of the RGB image-based HPE framework and the depth information-based HPE framework. We evaluated the performance of the proposed method by COCO Object Keypoint Similarity library. By taking an advantage of RGB image-based HPE method and depth information-based HPE method, our HPE method based on RGB-D achieved the mAP of 0.903 and mAR of 0.938. It proved that our method outperforms the RGB-based HPE and the depth-based HPE.

• Journal of the Korea Society of Computer and Information
• /
• v.21 no.10
• /
• pp.11-19
• /
• 2016

The size of a display is large, The form becoming various of that do not apply to previous methods of gaze tracking and if setup gaze-track-camera above display, can solve the problem of size or height of display. However, This method can not use of infrared illumination information of reflected cornea using previous methods. In this paper, Robust pupil detecting method for eye's occlusion, corner point of inner eye and center of pupil, and using the face pose information proposes a method for calculating the simply position of the gaze. In the proposed method, capture the frame for gaze tracking that according to position of person transform camera mode of wide or narrow angle. If detect the face exist in field of view(FOV) in wide mode of camera, transform narrow mode of camera calculating position of face. The frame captured in narrow mode of camera include gaze direction information of person in long distance. The method for calculating the gaze direction consist of face pose estimation and gaze direction calculating step. Face pose estimation is estimated by mapping between feature point of detected face and 3D model. To calculate gaze direction the first, perform ellipse detect using splitting from iris edge information of pupil and if occlusion of pupil, estimate position of pupil with deformable template. Then using center of pupil and corner point of inner eye, face pose information calculate gaze position at display. In the experiment, proposed gaze tracking algorithm in this paper solve the constraints that form of a display, to calculate effectively gaze direction of person in the long distance using single camera, demonstrate in experiments by distance.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.16 no.2
• /
• pp.131-139
• /
• 2016

Recognition of human motions has become a main area of computer vision due to its potential human-computer interface (HCI) and surveillance. Among those existing recognition techniques for human motions, head detection and tracking is basis for all human motion recognitions. Various approaches have been tried to detect and trace the position of human head in two-dimensional (2D) images precisely. However, it is still a challenging problem because the human appearance is too changeable by pose, and images are affected by illumination change. To enhance the performance of head detection and tracking, the real-time three-dimensional (3D) data acquisition sensors such as time-of-flight and Kinect depth sensor are recently used. In this paper, we propose an effective feature extraction method, called adaptive local binary pattern (ALBP), for depth image based applications. Contrasting to well-known conventional local binary pattern (LBP), the proposed ALBP cannot only extract shape information without texture in depth images, but also is invariant distance change in range images. We apply the proposed ALBP for head detection and tracking in depth images to show its effectiveness and its usefulness.

• Korean Journal of Head & Neck Oncology
• /
• v.26 no.1
• /
• pp.41-44
• /
• 2010

Hemangiomas are the most common benign tumors or tumor-like malformations of blood vessels or blood filled cavernous spaces. Hemangiomas of the oral cavity represents 14% of all hemangiomas. Most hemangiomas are detected by one year of age and most commonly occurred in the lips, tongue, buccal mucosa and palate. But hemangiomas are sometimes reported in adults. We report one case of extensive hemangioma of the tongue in adult. Our case is a 48-year-old male who presented with a hemangioma occupying entirely two-third of the tongue. Treating those lesions pose a challenge to the surgeon. This paper discusses one case of extensive tongue hemangiomas treated with alcoholic sclerotherapy with satisfactory results supporting this particular approach in the management of these lesion.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.51 no.6
• /
• pp.258-266
• /
• 2002

This paper proposes the methodology of the image processing algorithm that estimates the pose of the inner-pipe crawling robot. The inner-pipe crawling robot is usually equipped with a lighting device and a camera on its head for monitoring and inspection purpose of defects on the pipe wall and/or the maintenance operation. The proposed methodology is using these devices without introducing the extra sensors and is based on the fact that the position and the intensity of the reflected light from the inner wall of the pipe vary with the robot posture and the camera. The proposed algorithm is divided into two parts, estimating the translation and rotation angle of the camera, followed by the actual pose estimation of the robot . Based on the fact that the vanishing point of the reflected light moves into the opposite direction from the camera rotation, the camera rotation angle can be estimated. And, based on the fact that the most bright parts of the reflected light moves into the same direction with the camera translation, the camera position most bright parts of the reflected light moves into the same direction with the camera translation, the camera position can be obtained. To investigate the performance of the algorithm, the algorithm is applied to a sewage maintenance robot.

• Proceedings of the Korean Information Science Society Conference
• /
• 1999.10b
• /
• pp.401-403
• /
• 1999

본 논문에서 우리는 제약이 없는 배경화면에서 얼굴의 움직임을 이용한 응시점 추적을 위해 얼굴의 특징점(눈, 코, 그리고 입)들을 찾고 head orientation을 구하는 효？거이고 빠른 방법을 제안한다. 얼굴을 찾는 방법이 많이 연구 되어 오고 있으나 많은 부분이 효과적이지 못하거나 제한적인 사항을 필요로 한다. 본 논문에서 제안한 방법은 이진화된 이미지에 기초하고 완전 그래프 매칭을 이용한 유사성을 구하는 방법이다. 즉, 임의의 임계치 값에 의해 이진화된 이미지를 레이블링 한 후 각 쌍의 블록에 대한 유사성을 구한다. 이때 두 눈과 가장 유사성을 갖는 두 블록을 눈으로 선택한다. 눈을 찾은 후 입과 코를 찾아간다. 360$\times$240 이미지의 평균 처리 속도는 0.2초 이내이고 다음 탐색영역을 예상하여 탐색 영역을 줄일 경우 평균 처리속도는 0.15초 이내였다. 그리고 본 논문에서는 얼굴의 움직임을 구하기 위해 각 특징점들이 이루는 각을 기준으로 한 템플릿 매칭을 이용했다. 실험은 다양한 조명환경과 여러 사용자를 대상으로 이루어졌고 속도와 정확성면에서 좋은 결과를 보였다. 도한, 명안정보만을 사용하므로 흑백가메라에서도 사용가능하여 경제적 효과도 기대할 수 있다.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.173.1-173
• /
• 2001

This paper proposes the methodology in image processing algorithm that estimates the pose of the pipe crawling robot. The pipe crawling robots are usually equipped with a lighting device and a camera on its head for monitoring and inspection purpose. The proposed methodology is using these devices without introducing the extra sensors and is based on the fact that the position and the intensity of the reflected light varies with the robot posture. The algorithm is divided into two parts, estimating the translation and rotation angle of the camera, followed by the actual pose estimation of the robot. To investigate the performance of the algorithm, the algorithm is applied to a sewage maintenance robot.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2011.10a
• /
• pp.173-175
• /
• 2011

In this paper, we present an user interface which can show some 3D objects at various angles using tracked 3d head position and orientation. In implemented user interface, First, when user's head moves left/right (X-Axis) and up/down(Y-Axis), displayed objects are moved towards user's eyes using 3d head position. Second, when user's head rotate upon an X-Axis(pitch) or an Y-Axis(yaw), displayed objects are rotated by the same value as user's. The results of experiment from a variety of user's position and orientation show good accuracy and reactivity for 3d visualization.