• The Journal of Korea Robotics Society
• /
• v.17 no.3
• /
• pp.353-358
• /
• 2022

Due to the limited field of view of the pinhole camera, there is a lack of stability and accuracy in camera pose estimation applications such as visual SLAM. Nowadays, multiple-camera setups and large field of cameras are used to solve such issues. However, a multiple-camera system increases the computation complexity of the algorithm. Therefore, in multiple camera-assisted visual simultaneous localization and mapping (vSLAM) the multi-view tracking algorithm is proposed that can be used to balance the budget of the features in tracking and local mapping. The proposed algorithm is based on PanoSLAM architecture with a panoramic camera model. To avoid the scale issue 3D LiDAR is fused with omnidirectional camera setup. The depth is directly estimated from 3D LiDAR and the remaining features are triangulated from pose information. To validate the method, we collected a dataset from the outdoor environment and performed extensive experiments. The accuracy was measured by the absolute trajectory error which shows comparable robustness in various environments.

• Journal of Korea Multimedia Society
• /
• v.25 no.4
• /
• pp.575-582
• /
• 2022

In the manufacturing environment, research to minimize robot collisions with human beings have been widespread, but in order to interact with robots, it is important to precisely recognize and predict human actions. In this research, after enhancing performance by applying group normalization to the Hourglass model to detect the operator motion, the skeleton was estimated and data were created using this model. And then, three types of operator's movements were recognized using LSTM. As results of the experiment, the accuracy was enhanced by 1% using group normalization, and the recognition accuracy was 99.6%.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.47-48
• /
• 2021

본 논문에서는 대용량의 3D 데이터 시퀀스의 압축을 진행한다. 3D 데이터 시퀀스의 각 프레임에서 Pose Estimation을 통해 3D Skeleton을 추출한 뒤, 포인트 클라우드를 skeleton에 묶는 리깅 과정을 거치고, 다음 프레임과 같은 자세로 deformation을 진행한다. 다음 프레임과 같은 자세로 변형된 포인트 클라우드와 실제 다음 프레임의 포인트 클라우드를 비교하여, 두 데이터에 모두 있는 점, 실제 다음 프레임에만 있는 점, deformation한 데이터에만 있는 점으로 분류한다. 두 데이터에 모두 있는 점을 제외하고 나머지 두 분류의 점들을 저장함으로써 3D 시퀀스 데이터를 압축할 수 있다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.07a
• /
• pp.71-74
• /
• 2022

본 논문은 최적화알고리즘을 이용한 관절각 기반 3차원 자세 추정 기법의 정확도를 휴머노이드 로봇을 이용하여 검증하는 방법을 제안한다. 구글의 자세 추정 오픈소스 패키지인 MPP(MediaPipe Pose)로 특정자세를 취한 휴머노이드 로봇의 관절 좌표를 카메라의 픽셀 좌표로 추출한다. 추출한 픽셀 좌표를 전역최적화 방법인 uDEAS(univariate Dynamic Encoding Algorithm for Searches)를 통해 시상면과 관상면에서의 각도를 추정하고 휴머노이드 로봇의 실제 관절 각도와 비교하여 알고리즘의 정확도를 검증하는 방법을 제시한다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.5
• /
• pp.347-356
• /
• 2003

Camera pose and scene geometry estimation from video sequences is widely used in various areas such as image composition. Structure and motion recovery based on the auto calibration algorithm can insert synthetic 3D objects in real but un modeled scenes and create their views from the camera positions. However, most previous methods require bundle adjustment or non linear minimization process [or more precise results. This paper presents a new auto' calibration algorithm for video sequence based on two steps: the one is key frame selection, and the other removes the key frame with inaccurate camera matrix based on an absolute quadric estimation by LMedS. In the experimental results, we have demonstrated that the proposed method can achieve a precise camera pose estimation and scene geometry recovery without bundle adjustment. In addition, virtual objects have been inserted in the real images by using the camera trajectories.

• Journal of Biomedical Engineering Research
• /
• v.44 no.3
• /
• pp.167-175
• /
• 2023

Chest PA is the basic examination of radiographic imaging. Moreover, Chest PA's demands are constantly increasing because of the Increase in respiratory diseases. However, it is not meeting the demand due to problems such as a shortage of radiological technologist, sexual shame caused by patient contact, and the spread of infectious diseases. There have been many cases of using artificial intelligence to solve this problem. Therefore, the purpose of this research is to build an artificial intelligence dataset of Chest PA and to find a posture evaluation method. To construct the posture dataset, the posture image is acquired during actual and simulated examination and classified correct and incorrect posture of the patient. And to evaluate the artificial intelligence posture method, a posture estimation algorithm is used to preprocess the dataset and an artificial intelligence classification algorithm is applied. As a result, Chest PA posture dataset is validated with in over 95% accuracy in all artificial intelligence classification and the accuracy is improved through the Top-Down posture estimation algorithm AlphaPose and the classification InceptionV3 algorithm. Based on this, it will be possible to build a non-face-to-face automatic Chest PA examination system using artificial intelligence.

• Journal of Electrical Engineering and Technology
• /
• v.11 no.2
• /
• pp.508-518
• /
• 2016

In this paper, we propose a method combining an accelerometer with a cross structured light system to estimate the golf green slope. The cross-line laser provides two laser planes whose functions are computed with respect to the camera coordinate frame using a least square optimization. By capturing the projections of the cross-line laser on the golf slope in a static pose using a camera, two 3D curves’ functions are approximated as high order polynomials corresponding to the camera coordinate frame. Curves’ functions are then expressed in the world coordinate frame utilizing a rotation matrix that is estimated based on the accelerometer’s output. The curves provide some important information of the green such as the height and the slope’s angle. The curves estimation accuracy is verified via some experiments which use OptiTrack camera system as a ground-truth reference.

• Journal of Broadcast Engineering
• /
• v.25 no.5
• /
• pp.750-757
• /
• 2020

One of the basic tasks for robots to interact with humans is to quickly and accurately grasp human behavior. Therefore, it is necessary to increase the accuracy of human pose recognition when the robot is estimating the human pose and to recognize it as quickly as possible. However, when the human pose is estimated using deep learning, which is a representative method of artificial intelligence technology, recognition accuracy and speed are not satisfied at the same time. Therefore, it is common to select one of a top-down method that has high inference accuracy or a bottom-up method that has high processing speed. In this paper, we propose two methods that complement the disadvantages while including both the advantages of the two methods mentioned above. The first is to perform parallel inference on the server using multi GPU, and the second is to mix bottom-up and One-class Classification. As a result of the experiment, both of the methods presented in this paper showed improvement in speed. If these two methods are applied to the entertainment robot, it is expected that a highly reliable interaction with the audience can be performed.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.42 no.4 s.304
• /
• pp.11-18
• /
• 2005

For expressing intention the human often use body languages as well as vocal languages. Of course the gestures using arms and hands are the representative ones among the body languages. Therefore it is very important to understand the human arm motion in human-computer interaction. In this respect we present here how to estimate 3D pose of human arms by using computer vision systems. For this we first focus on the idea that the human arm motion consists of mostly revolute joint motions, and then we present an algorithm for understanding 3D motion of a revolute joint using vision systems. Next we apply it to estimating 3D pose of human arms using vision systems. The fundamental idea for this algorithm extension is that we may apply the algorithm for a revolute joint to each of the revolute joints of hmm arms one after another. In designing the algorithms we focus on seeking closed-form solutions with high accuracy because we aim at applying them to human computer interaction for ubiquitous computing and virtual reality.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.5
• /
• pp.315-324
• /
• 2023

In this paper, we propose a real-time pose estimation method for a monocular camera using a particle filter integrated with UKF (unscented Kalman filter). While conventional camera tracking techniques combine camera images with data from additional devices such as gyroscopes and accelerometers, the proposed method aims to use only two-dimensional visual information from the camera without additional sensors. This leads to a significant simplification in the hardware configuration. The proposed approach is based on a particle filter integrated with UKF. The pose of the camera is estimated using UKF, which is defined individually for each particle. Statistics regarding the camera state are derived from all particles of the particle filter, from which the real-time camera pose information is computed. The proposed method demonstrates robust tracking, even in the case of rapid camera shakes and severe scene occlusions. The experiments show that our method remains robust even when most of the feature points in the image are obscured. In addition, we verify that when the number of particles is 35, the processing time per frame is approximately 25ms, which confirms that there are no issues with real-time processing.