Myo-Taeg Lim

Korea University, Sŏul, Seoul, South Korea

Are you Myo-Taeg Lim?

Claim your profile

Publications (73)19.88 Total impact

  • Yan-Feng Lu · Myo-Taeg Lim · Hua-Zhen Zhang · Tae-Koo Kang
    IET Computer Vision 06/2015; DOI:10.1049/iet-cvi.2014.0249 · 0.76 Impact Factor
  • Transactions of the Korean Institute of Electrical Engineers 05/2015; 64(5):779-785. DOI:10.5370/KIEE.2015.64.5.779
  • [Show abstract] [Hide abstract]
    ABSTRACT: Hierarchical Model and X (HMAX) presents a biologically inspired model for robust object recognition. The HMAX model, based on the mechanisms of the visual cortex, can be described as a four-layer structure. Although the performance of HMAX in object recognition is robust, it has been shown to be sensitive to rotation, which limits the model׳s performance. To alleviate this limitation, we propose an Oriented Gaussian–Hermite Moment-based HMAX (OGHM-HMAX). In contrast to HMAX which uses a Gabor filter for local feature representation, OGHM-HMAX employs the Oriented Gaussian–Hermite Moment (OGHM), which is a local representation method that represents features and is robust against distortions. OGHM is an extension of the modified discrete Gaussian–Hermite moment (MDGHM). To show the effectiveness of the proposed method, experimental studies on object categorization are conducted on the CalTech101, CalTech5, Scene13 and GRAZ01 databases. Experimental results demonstrate that the performance of OGHM-HMAX is a significant improvement on that of the conventional HMAX.
    Neurocomputing 09/2014; 139:189–201. DOI:10.1016/j.neucom.2014.02.046 · 2.01 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: At high-rise building construction sites, construction workers still climb dangerous vertical steel beams. Hence, the bolting cabin helps construction workers to assemble bolts safely. In order to automatically assemble bolts to steel frame, the bolting cabin must estimate bolt-hole location. To do this, precise bolt-hole detection algorithms are important. Generally, a 2D camera is used for visual servoing in industrial fields. However, a construction site is an outdoor environment that does not have constant illumination. Recently, a 3D camera that provides depth images has been used in industrial fields. However, this 3D camera provides a low resolution image with strong noise around the edges. To overcome these problems, an edge detection filter that is robust to strong noise is needed for precise bolt-hole detection. Therefore, we propose a Modified Discrete Gaussian-Hermite Moments (MDGHM) filter based on moment information to detect edges precisely. Using the MDGHM filter, the results show that the proposed method can be effectively applied to bolt-hole detection better than wavelet filter and mixed matching method.
    Automation in Construction 08/2014; 44:1–11. DOI:10.1016/j.autcon.2014.03.022 · 1.82 Impact Factor
  • Tae-Koo Kang · In-Hwan Choi · Myo-Taeg Lim
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes a novel family of local feature descriptors, a variant of the speed up robust features (SURF) descriptor, which is capable of demonstrably better performance. The conventional SURF descriptor is an efficient implementation of the SIFT descriptor. Although the SURF descriptor can represent the nature of the underlying image pattern, it is still sensitive to more complicated deformations such as large viewpoint and rotation changes. To solve this problem, our family of descriptors, called MDGHM-SURF, is based on the modified discrete Gaussian–Hermite moment (MDGHM), which devises a movable mask to represent the local feature information of non-square images. Whereas conventional SURF uses first-order derivatives, MDGHM-SURF uses MDGHM, which offers more feature information than first-order derivative-based local descriptors such as SURF and SIFT. Consequently, by redefining the conventional SURF descriptor using MDGHM, MDGHM-SURF can extract more distinctive features than conventional SURF. The results of evaluations conducted with six types of deformations indicate that our proposed method outperforms the matching accuracy of other SURF related algorithms.
    Pattern Recognition 07/2014; 48(3). DOI:10.1016/j.patcog.2014.06.022 · 2.58 Impact Factor
  • Tae-Koo Kang · Myo-Taeg Lim · Gwi-Tae Park · Dong W. Kim
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper addresses the vision based local path planning system for obstacle avoidance. To handle the obstacles which exist beyond the field of view (FOV), we propose a Panoramic Environment Map (PEM) using the MDGHM-SIFT algorithm. Moreover, we propose a Complexity Measure (CM) and Fuzzy logic-based Avoidance Motion Selection (FAMS) system to enable a humanoid robot to automatically decide its own direction and walking motion when avoiding an obstacle. The CM provides automation in deciding the direction of avoidance, whereas the FAMS system chooses the avoidance path and walking motion, based on environment conditions such as the size of the obstacle and the available space around it. The proposed system was applied to a humanoid robot that we designed. The results of the experiment show that the proposed method can be effectively applied to decide the avoidance direction and the walking motion of a humanoid robot.
    Journal of Electrical Engineering and Technology 07/2013; 8(4):879-888. DOI:10.5370/JEET.2013.8.4.879 · 0.52 Impact Factor
  • Baeksuk Chu · Kyoungmo Jung · Myo-Taeg Lim · Daehie Hong
    [Show abstract] [Hide abstract]
    ABSTRACT: The construction industry has not traditionally been a favorable field for the application of robotic technologies. However, various motivations such as the shrinking labor population, the aging of skilled workers, and the safety issue of ironworkers have promoted the development of robotic construction systems. In this research, one of those trials, a project entitled “Robot-based construction automation system for high-rise building” is presented. Among diverse construction works, this project focused on a robotic automation of the steel beam assembly. The project is a cooperative effort between a robot research group and a construction automation group in South Korea. The main objective of this paper is to an introduction for the development of a robotic beam assembly system administered by the robot research group. The robotic beam assembly system consists of a robotic bolting device that performs the main function for the beam assembly work and a robotic transport mechanism that transports the robotic bolting device to target bolting positions around a building under construction. This paper presents the specific functions, structures, and mechanisms of the robotic bolting device and accounts for the application of the visual servo control technique to a bolting control system which is a software component. The robotic transport mechanism part is discussed in a companion paper [14]. The real prototype of the proposed system was manufactured and intensive field tests were conducted in a test bed. Moreover, this system was applied to a section of a real building, the Robot Convergence Building of Korea University, South Korea, which has one story below and seven above the ground, and obtained a feasibility of an application of the robotic beam assembly system to actual construction sites. The suggested system is expected to be a promising alternative to ironworkers in the steel beam assembly in terms of safety and time-efficiency.
    Automation in Construction 07/2013; 32:46–61. DOI:10.1016/j.autcon.2012.12.016 · 1.82 Impact Factor
  • Hyun-Duck Choi · Choon-Ki Ahn · Myo-Taeg Lim
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes a new resampling algorithm - Gaussian distributed resampling (GDR) algorithm - in order to solve particle impoverishment phenomenon in particle filter. The key idea of proposed algorithm generates the new particles based on Gaussian distribution, which depend on the size of weight in resampling process. In comparison with established resampling algorithms, diversity of particles can be maintained, and thus the proposed algorithm avoids the sample impoverishment.
    Control, Automation and Systems (ICCAS), 2013 13th International Conference on; 01/2013
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents obstacle avoidance method for scout robot or industrial robot in unknown environment by using IR sensor and vision system. In the proposed method, robots share the information where the obstacles are located in real-time, thus the robots can choose the best path for obstacle avoidance. Using IR sensor and vision system, multiple robots efficiently evade the obstacles by the proposed cooperation method. No landmark is used at wall or floor in experiment environment. The obstacles don't have specific color or shape. To get the information of the obstacle, vision system extracts the obstacle coordinate by using an image labeling method. The information obtained by IR sensor is about the obstacle range and the locomotion direction to decide the optimal path for avoiding obstacle. The experiment was conducted in indoor environment with two-wheeled mobile robots. It is shown that multiple robots efficiently move along the optimal path in cooperation with each other in the space where obstacles are located.
    Journal of Institute of Control 12/2012; 18(12). DOI:10.5302/J.ICROS.2012.18.12.1122
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Three-dimensional (3D) imaging is gaining popularity and has been partly adopted in laparoscopic surgery or robotic surgery but has not been applied to gastrointestinal endoscopy. As a first step, we conducted an experiment to evaluate whether images obtained by conventional gastrointestinal endoscopy could be used to acquire quantitative 3D information. Two endoscopes (GIF-H260) were used in a Borrmann type I tumor model made of clay. The endoscopes were calibrated by correcting the barrel distortion and perspective distortion. Obtained images were converted to gray-level image, and the characteristics of the images were obtained by edge detection. Finally, data on 3D parameters were measured by using epipolar geometry, two view geometry, and pinhole camera model. The focal length (f) of endoscope at 30 mm was 258.49 pixels. Two endoscopes were fixed at predetermined distance, 12 mm (d(12)). After matching and calculating disparity (v2-v1), which was 106 pixels, the calculated length between the camera and object (L) was 29.26 mm. The height of the object projected onto the image (h) was then applied to the pinhole camera model, and the result of H (height and width) was 38.21 mm and 41.72 mm, respectively. Measurements were conducted from 2 different locations. The measurement errors ranged from 2.98% to 7.00% with the current Borrmann type I tumor model. It was feasible to obtain parameters necessary for 3D analysis and to apply the data to epipolar geometry with conventional gastrointestinal endoscope to calculate the size of an object.
    Clinical Endoscopy 09/2012; 45(3):182-8. DOI:10.5946/ce.2012.45.3.182
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, a 3D vision-based local obstacle avoidance system is designed and developed on a humanoid robot so that it can decide avoidance direction and walking motion effectively. We use a panorama environment map using speeded up robust feature (SURF) which is a robust image detector and descriptor to handle the obstacles which exist beyond the field of view. Moreover, we propose an avoidance direction decision method and a fuzzy logic based avoidance motion selection method. The robot decides the avoidance direction and avoidance walking motion for the obstacle by itself under information such as the size of objects and avoidance spaces. The proposed system is applied to the humanoid robot which we have built up with a Time of Flight camera. The results of the experiments show that the proposed method can be effectively applied to decide the avoidance direction and the walking motion.
    Control, Automation and Systems (ICCAS), 2012 12th International Conference on; 01/2012
  • International Journal of Advanced Robotic Systems 01/2012; DOI:10.5772/50980 · 0.50 Impact Factor
  • Eun-Hye Kim · Myo-Taeg Lim · Yong-Kwun Lee
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents the analysis of grasp stability for a bio-mimetic robot hand by using polygon segmentation algorithm. The presented hand has four independency moving fingers, which are driven by four-coupled link mechanism with two linear actuators. The robot hand was designed considering the dexterity and the compact size suited for various objects in daily life. Tactile sensors are mounted on the palm and the fingertip at each finger of the hand. With tactile sensor based on feedback control, the paper explains the control of the hand which is very stable in grasping by using new algorithm called polygon segmentation algorithm. This algorithm can be appiled for the analysis of grasp stability to various objects regardless of the number of contact points. Using the force equilibrium points and vertex of the polygon, several triangles can be obtained. We can obtain the polygon segmentation algorithm by analyzing the relationship among triangles. In order to verify this algorithm, experiments were conducted by using three types of objects with different characteristics.
  • Yung-Hak Mo · Jong-Wook Lim · Jung-Min Park · Myo-Taeg Lim
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper suggest a method to detect the bolt hole of the steel frame using the new detection that is combined template matching and the circular Hough transform. The template matching can be used to roughly estimate the location of the object under the various illumination. The circular Hough transform is used to discover more accurate location of the object than the template matching. Therefore, we compare the performance between each individual method and the mixed methods. We confirm that the proposed method is robust with respect to illumination.
    01/2011;
  • [Show abstract] [Hide abstract]
    ABSTRACT: The Bolting robot assistance system prevents operator to face dangerous situations. Bolting robot assistance system consists of bolting robot control system and top-view supervisory system. In order to control a bolting robot, Camshift algorithm and circular Hough transform are used. To estimate location of bolt hole, circular Hough transform is used. In order to track a bolt hole, Meanshift and Camshift are used. Camshift operates on color image represented by probability point and applies a non-parametric gradient density called Meanshift algorithm to re-center its operating window. By setting region to search a bolt hole, the algorithm will track the location of the bolt hole. In order to make top-view supervisory system, four cameras are installed at left, right, front and back of the robot. Each image from the camera is used to make the top-view image after correcting distortion. This paper proposes the image processing algorithm which is suitable for top-view supervisory system.
    Control Automation and Systems (ICCAS), 2010 International Conference on; 11/2010
  • Sang-Hyuk Park · Young-Joong Kim · Myo-Taeg Lim
    [Show abstract] [Hide abstract]
    ABSTRACT: Precise estimation of the position of robots, which is essential in mobile robotics, is difficult to achieve. However, particle filter shows great promise in this area. The number of samples used in this study is closely related to the operation time in particle filtering. The main issue in real-time implementation with regard to particle filter is to reduce the operation time, which led to the development of the adaptive particle filter (APF). We propose a new APF which adjusts the variance and then uses the gradient data to generate samples near the high likelihood region. The experiment results show that the new APF performs better, in terms of the total operation time and sample set size, than the standard particle filter and the APF using Kullback-Leibler distance sampling. KeywordsKullback-leibler distance-mobile robot-particle filter-ultrasonic beacon
    International Journal of Control Automation and Systems 08/2010; 8(4):801-807. DOI:10.1007/s12555-010-0412-4 · 1.07 Impact Factor
  • Eun-Hye Kim · Myo-Taeg Lim · Yong-Kwun Lee
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper describes grasp stability for a multi-fingered robot hand, called the KIST hand. The robot hand has controlled by four under-actuated fingers with totally nine DOFs, which are controlled by two linear actuators and linkage knuckles. This mechanism is able to generate high power compare to common robot hands that use rotary actuators. The robot hand has four tactile sensors which are attached to the fingertips on the each finger. The each tactile sensor can independently measure the contact force between the robot hand and an object. Using the tactile sensors, various objects can be grasped, such as a small tennis ball, a plastic ball and a rubber ball. In the former part of the paper, the mechanical design of the robot hand is presented. In the latter part of the paper, the control algorithm is described, and the analysis of grasp stability is confirmed by the experimental result.
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper we will use stereo camera for 3D modeling of object in indoor environment. For mobile robot, navigation is a very important research area. Stereo vision can provide many information of the environment, and with these information robot can perform high level task. With our stereo vision system, firstly, we get disparity map from frames captured by the stereo camera. Then we extract the points of object from the frames, and convert these points to 3D coordinates. We select key points from the object points. Finally with these key points we build D model by our modeling system. Then we can get a virtual model of the objects in indoor environment.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Recently, a development of the medical instrument using the vision information is brisk. Especially, extracting the 3-dimension information from 2-dimension image is the one of the major research topics. This paper proposes the method to measure tumor size by the 3-dimension information extraction, the triangulation using the extracted 3-dimension information and the camera geometry. To extract the 3-dimension information, the Hough circle transform and triangulation are used. The Hough circle transform is used to extract a feature point from tumor models for the 3-dimension information recovery. An extracted feature point is used for recovering the depth information from the 2-dimension tumor stereo images. A tumor size measurement is done by triangulation using the depth information and camera internal parameters. Matlab and Visual C++ are used for simulation.
  • [Show abstract] [Hide abstract]
    ABSTRACT: In order to use a robot in the construction automation filed, we proposed the concept of Bolting Robot and a visual servo control scheme to track a bolting tool to a bolt hole in the structural steel frame. For estimating a location of a bolt hole, Circular Hough Transform (CHT) was used to extract circles. Generally, CHT is computationally complex due to a power of the dimensionality of a circle. A distance from a camera to a steel frame can be measured by using laser range-finder installed. The radius of a bolt hole can be calculated with the distance to a steel frame. Since the radius is known, the processing of CHT can be reduced to 2D. In addition, it contains image pre-processing to make an image of bolt holes to be clear. Pre-processing has 4 steps which consist of compensating lens distortion, noise filtering, histogram equalization, and edge detection.
    ICCAS-SICE, 2009; 09/2009