http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Parallel optimal control with multiple shooting, constraintsaggregation and adjoint methods
Moongu Jeon 한국전산응용수학회 2005 Journal of applied mathematics & informatics Vol.19 No.1-2
In this paper, constraint aggregation is combined with the adjoint and multiple shooting strategies for optimal control of differential algebraic equations (DAE) systems. The approach retains the inherent parallelism of the conventional multiple shooting method, while also being much more efficient for large scale problems. Constraint aggregation is employed to reduce the number of nonlinear continuity constraints in each multiple shooting interval, and its derivatives are computed by the adjoint DAE solver DASPKADJOINT together with ADIFOR and TAMC, the automatic differentiation software for forward and reverse mode, respectively. Numerical experiments demonstrate the effectiveness of the approach.
Learning-based essential matrix estimation for visual localization
Son Moongu,Ko Kwang-Hee 한국CDE학회 2022 Journal of computational design and engineering Vol.9 No.3
Visual localization is defined as finding the camera pose from two-dimensional images, which is a core technique in many computer vision tasks, including robot navigation, autonomous driving, augmented/mixed/virtual reality, mapping, etc. In this study, we address the pose estimation problem from a single-color image using a neural network. We propose a coarse-to-fine approach based on a deep learning framework, which consists of two steps: direct regression-based coarse pose estimation that obtains a pose by finding a pose-based similar image retrieval and Siamese network-based essential matrix estimation to obtain a refined pose. Experimental results using the 7-scenes, Cambridge, and RobotCar datasets demonstrate that the proposed method performs better than the existing methods in terms of accuracy and stability.
손문구(MoonGu Son),이관행(Kwan H Lee) (사)한국CDE학회 2016 한국 CAD/CAM 학회 학술발표회 논문집 Vol.2016 No.동계
Deformation lamps is technique to adding moving effect on static object by projecting luminance pattern unlike conventional projection mapping technique that modify the appearance of object. It is necessary that deformed image sequence shows movement. However these deformed image sequences made by manual or video that has movement effect. In this paper, using motion texture creates movement effect from target object’s still image. This technique has given the dynamic effect of the static object causing the interest of the audience and can be applied to advertising, interior design, art, and exhibition.
Seokhyoung Lee,Moongu Jeon,Shin, V. IEEE 2012 IEEE transactions on industrial electronics Vol.59 No.11
<P>A new distributed fusion filtering algorithm for linear multiple time-delayed systems is proposed. The multisensory distributed fusion filter is formed by the summation of local Kalman filters having time delays (LKFTDs) in both the system and measurement models. The proposed distributed filter has a parallel structure that enables processing of multisensory measurements; thereby, it is more reliable than the centralized version if some sensors turn faulty. The key contribution of this paper is the derivation of recursive error cross-covariance equations between the LKFTDs to compute the optimal matrix fusion weights. In the particular case of multisensory dynamic systems having time delays in only the measurement model, the obtained results coincide with the previous work of Sun. The high accuracy and efficiency of the proposed distributed filter are then demonstrated through its implementation on a vehicle suspension system.</P>
( Sergey Yun ),( Moongu Jeon ) 한국정보처리학회 2014 한국정보처리학회 학술대회논문집 Vol.21 No.1
In this work we present a robust and fast approach to estimate 3D vehicle pose that can provide results under a specific traffic surveillance conditions. Such limitations are expressed by single fixed CCTV camera that is located relatively high above the ground, its pitch axes is parallel to the reference plane and the camera focus assumed to be known. The benefit of our framework that it does not require prior training, camera calibration and does not heavily rely on 3D model shape as most common technics do. Also it deals with a bad shape condition of the objects as we focused on low resolution surveillance scenes. Pose estimation task is presented as PnP problem to solve it we use well known “POSIT” algorithm [1]. In order to use this algorithm at least 4 non coplanar point’s correspondence is required. To find such we propose a set of techniques based on model and scene geometry. Our framework can be applied in real time video sequence. Results for estimated vehicle pose are shown in real image scene.
Particle Filtering과 계층적인 Boosting 알고리즘을 기반으로 한 다중 객체 추적 연구
양이화(Ehwa Yang),전문구(Moongu Jeon) 한국정보과학회 2012 한국정보과학회 학술발표논문집 Vol.39 No.1B
본 논문은 Particle Filtering과 계층적인 Boosting 알고리즙을 이용한 다중 객체 추적 기법을 제안한다. Particle Filtering 을 이용하여 각 객체를 단일 객체로 추적하고 Boosting 기반의 데이터 연관 알고리즘을 사용하여 영상에서 움직이는 물체들을 추적한다. 본 제안한 알고리즘에서는 객체들의 이동경로 정확한 감지를 위해 Particle Filtering을 통해 각 객체가 움직이는 예측 정보를 이용하고, Boosting 알고리즘을 계층적인 형태로 설계함에 따라 데이터 물체의 추적 정확도를 놓일 수 있도록 하였다.
Global Abnormal Event Detection using Spatio-Temporal Social Force Model
Jongmin Yu,Jeonghwan Gwak,Moongu Jeon 대한전자공학회 2015 대한전자공학회 학술대회 Vol.2015 No.6
In this work, we propose the spatio-temporal social force model (STSFM) for global abnormal event (GAE) detection in crowded scenes. While STSFM is based on the social force model (SFM), it is further improved by incorporating a temporal property of moving objects from consecutive frame sequences. In the proposed method, the interaction force of STSFM is then mapped into not only single frame plane but also frame sequences to obtain the force flow for every pixel in each frame. The experiments were conducted on the UMN dataset from University of Minnesota. The results clearly demonstrated the significance of STSFM by showing that it outperforms similar approaches based on the pure optical flow and SFM in terms of both quantitative and qualitative measures.