Won-Ho Choi

University of Ulsan, Ulsan, Ulsan, South Korea

Are you Won-Ho Choi?

Claim your profile

Publications (8)1.24 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: We propose a method for detection and tracking for objects under multiple cameras system. To track objects, one need to establish correspondence objects among multiple views. We apply the principal axis of objects and the homography constraint to match objects across multiple cameras. The principal axis belongs to the silhouette of objects that is extracted by the background subtraction. We use the multiple background model to the background subtraction. In an image sequence, many changes happen with respect to pixel intensity. This cannot be characterized by the single background model so that is necessary to use the multiple background model. Also, we use the median background model reducing some noises. The silhouette is detected by difference with background models and current image which includes moving objects. For calculating homography, we use landmarks on the ground plane in 3D space. The homography means the relation between two correspondence between two coinciding points from different views. The intersection of principal axes and ground plane in 3D space are the same point shown in each view. The intersection occurs when a principal axis in an image crosses to the transformed ground plane from another image. We construct the correspondence which means the relationship between intersection in current image and transformed intersection from the other image by homography constraint. Those correspondences confirm within a short distance measuring in the top viewed plane. Thus, we track a person by these corresponding points on the ground plane.
    No preview · Article · Jan 2009 · JOURNAL OF UNIVERSAL COMPUTER SCIENCE
  • Source
    Hang-Ki Ryu · Jae-Kook Lee · Byeong-Woo Kim · Won-Ho Choi
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we propose a new 3-D working point determination method for industrial robot using vision camera system and block interpolation technique with feature points in a vehicle body. To detect the feature points in a vehicle body, we applied the pattern matching method. For determination of working point, we applied block interpolation method. The block consists of 3-D type blocks with detected feature points per section. 3-D position is selected by Euclidean distance between 245 feature values and an acquired feature point. In order to evaluate the proposed algorithm, experiments are performed in glass equipment process in real industrial vehicle assembly line.
    Preview · Article · Jan 2008 · Transactions of the Korean Institute of Electrical Engineers
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a study on direct torque control method (DTC) using matrix converter fed induction motor. The advantages of matrix converters are combined with the advantages of the DTC technique: under the constraint of unity input power factor, the required voltage vectors are generated to implement the conventional DTC method of induction motor. However, since the first idea about DTC method using MC fed induction motor was suggested by the current researches just focus on some simulation results and a little un-explicit experimental result is carried out. This paper describes the operation of induction motor under the DTC method in steady-state and transient conditions by the experimental results, the discussion about the trend of DTC method using MC is also carried out. Furthermore, the entire system of matrix converter configuration using 7.5 kW IGBT modules (FR35R12KE3V1) is explained in detail.
    Preview · Conference Paper · Nov 2007
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes a new corner detection method based on the Hessian matrix. The proposed method can detect features of a pattern or input image using eigenvalue and eigenvector of images. The Hessian matrix has information of ellipse with intensity variance, and corner can be detected by using the eigen-value and eigen-vector analysis and decided weight value. In order to evaluate the proposed algorithm, experiments are performed in many type images. As the result of the test image, it shows the better performance than that of conventional Harris, SUSAN, and symmetric corner detectors.
    No preview · Article · Oct 2007
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes a self-calibration method of robots those are use in industrial vehicle assembly lines. The proposed method is a position compensation using laser sensor and vision cameras. Because the laser sensor is cross type sensor which can scan a horizontal and vertical line, it is efficient way to detect a feature of vehicle and winding shape of vehicle's body. For position compensation of 3-Dimensional axis, we applied block interpolation method. For selecting feature point, pattern matching method is used and 3-D position is selected by Euclidean distance mapping between 462 feature values and evaluated feature point. In order to evaluate the proposed algorithm, experiments are performed in real industrial vehicle assembly line. As the result of the field test, it shows that robot's working point can be displayed 3-D points. These points are used to diagnosis error of position and reselecting working point.
    Preview · Conference Paper · Nov 2006
  • Jae-Kuk Lee · Young-Jun Joung · Won-Ho Choi
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we propose a new classification method using local probability and statistical hypothesis theory. To separate the test data, we analyze the local area of the test data set using local probability distribution and decide the candidate class of the data set. To decide each class of the test data, statistical hypothesis theory is applied to the decided candidate class of the test data set. To evaluate, the proposed classification method is compared to the conventional fuzzy c-mean method and k-means algorithm. The simulation results show more accuracy than results of fuzzy c-mean method and k-means algorithm.
    No preview · Conference Paper · Jan 2004
  • Jae-Kuk Lee · Kyung-Hun Kim · Tae-Young Kim · Won-Ho Choi
    [Show abstract] [Hide abstract]
    ABSTRACT: Principle component analysis (PCA) is a dimensionality reduction technique for data analysis and processing, and it is a linear method that can maximize the variance of data. In this paper, to achieve the high efficiency for data classification, nonlinear PCA using local probability is proposed. Parameters are extracted from each distribution of data and mapping function of data set is made using the relation of the extracted parameters. Nonlinear PCA is performed in new projection feature space. The experimental result is conducted to verify its efficiency compared with the classical linear PCA.
    No preview · Conference Paper · Jan 2003
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a new data classification method using the Robbins-Monro stochastic approximation algorithm, k-nearest neighbor algorithm and probability analysis. The centroid of the test data set is decided by k-nearest neighbor algorithm and Robbins-Monro stochastic approximation algorithm. To decide the members of each class, the probability analysis is applied on the basis of the decided centroid point in data set. The proposed classification method is compared to the conventional fuzzy c-mean method, k-nn algorithm and discriminant analysis algorithm. The results show that the proposed method is more accurate than fuzzy c-mean method, k-means algorithm and discriminant analysis algorithm.
    Preview · Article ·