• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.3 s.303
• /
• pp.15-22
• /
• 2005

This paper proposed a object tracking algorithm which can track contour of fast moving object through motion estimation. Since the proposed tracking algorithm is based on the radial representation, the motion estimation of object can be accomplished at the center of object with the low computation complexity. The motion estimation of object makes it possible to track object which move fast more than distance from center point to contour point for each frame. In addition, by introducing both gradient image and difference image into energy functions in the process of energy convergence, object tracking is more robust to the complex background. The results of experiment show that the proposed algorithm can track fast moving object in real-time and is robust under the complex background.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.37 no.1
• /
• pp.32-42
• /
• 2000

This paper presents an on-line learning algorithm for sequential construction of generalized radial basis function networks (GRBFNs) to model nonlinear systems from empirical data. The GRBFN, an extended from of standard radial basis function (RBF) networks with constant weights, is an architecture capable of representing nonlinear systems by smoothly integrating local linear models. The proposed learning algorithm has a two-stage learning scheme that performs both structure learning and parameter learning. The structure learning stage constructs the GRBFN model using two construction criteria, based on both training error criterion and Mahalanobis distance criterion, to assign new hidden units and the linear local models for given empirical training data. In the parameter learning stage the network parameters are updated using the gradient descent rule. To evaluate the modeling performance of the proposed algorithm, simulations and their results applied to two well-known benchmarks are discussed.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.10
• /
• pp.2441-2449
• /
• 2015

In this digital information era, utilization of images are essential for various media, and the edge is an important characteristical information of an object in images that includes the size, location, direction and etc. Many domestic and international studies are being conducted in order to detect these edge. Existing edge detection methods include Sobel, Prewitt, Roberts, Laplacian, LoG and etc. which apply fixed weight value. As these existing edge detection methods apply fixed weight mask to the image, edge detection characteristic appears slightly insufficient. Accordingly, in order to supplement these problems, this study used bottom-hat transformation from mathematical morphology and opening operation in improving the image and proposed an algorithm that detects for the edge after calculating mask-based gradient. And to evaluate the performance of the proposed algorithm, a comparison was made against the existing Sobel, Roberts, Prewitt, Laplacian, LoG edge detection methods, in illustrating visual images, and similarities were compared by calculating the MSE value based on the standard of each image.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.7
• /
• pp.893-898
• /
• 2005

Due to the merits that only a small amount of computation is needed for solutions and stochastic policies can be handled explicitly, the actor-critic algorithm, which is a class of reinforcement learning methods, has recently attracted a lot of interests in the area of artificial intelligence. The actor-critic network composes of tile actor network for selecting control inputs and the critic network for estimating value functions, and in its training stage, the actor and critic networks take the strategy, of changing their parameters adaptively in order to select excellent control inputs and yield accurate approximation for value functions as fast as possible. In this paper, we consider a new actor-critic algorithm employing an RLS(Recursive Least Square) method for critic learning, and policy gradients for actor learning. The applicability of the considered algorithm is illustrated with experiments on the two linked robot arm.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.39 no.5
• /
• pp.433-441
• /
• 2011

The optimization of a centrifugal compressor was conducted. The ANN (Artificial Neural Network) was adopted as an optimization algorithm, and it was learned and trained with the DOE (Design of Experiment). In the DOE, it was predicted the main effect and the interaction effect of design variables to the objective function. The ANN was improved in the optimization process using the GA (Genetic Algorithm). When any output at each generation was reached a standard level, it was re-calculated by the CFD (Computational Fluid Dynamics) and it was applied to develop a new ANN. After 6th generation, the prediction difference between ANN and CFD was less than 1%. A pareto of the efficiency versus the pressure ratio was obtained through the 21th generation. Using this method, the computational time for the optimization was equivalent to the time consumed by the gradient method, and the optimized results of multi-objective function were obtained.

• Geophysics and Geophysical Exploration
• /
• v.22 no.4
• /
• pp.195-201
• /
• 2019

The logarithmic objective function used in waveform inversion minimizes the logarithmic differences between the observed and modeled data. Laplace-domain waveform inversions usually adopt the logarithmic objective function and the diagonal elements of the pseudo-Hessian for optimization. In this case, we apply the Levenberg-Marquardt algorithm to prevent the diagonal elements of the pseudo-Hessian from being zero or near-zero values. In this study, we analyzed the diagonal elements of the pseudo-Hessian of the logarithmic objective function and showed that there is no zero or near-zero value in the diagonal elements of the pseudo-Hessian for acoustic waveform inversion in the Laplace domain. Accordingly, we do not need to apply the Levenberg-Marquardt algorithm when we regularize the gradient direction using the pseudo-Hessian of the logarithmic objective function. Numerical examples using synthetic and field datasets demonstrate that we can obtain inversion results without applying the Levenberg-Marquardt method.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.5
• /
• pp.1051-1058
• /
• 2018

This paper propose image processing algorithm to prevent safety accidents near by subway platform screen door(PSD). First, edges of the subway PSD images are detected and the boundary line between PSD and subway platform is detected to decide people's approach to the PSD using Hough transform. To do this, we draw the boundary line between the PSD and platform, to detect the boundary line and to decide the people's approach to the detected line is completely connected or not. Generally, edge is the basic characteristic of image; thus, edge detection is very important in image processing applications and computer vision area. The conventional edge detection methods such as Roberts, Sobel, Prewitt, and Laplacian etc, which are using a fixed value of mask, and morphological gradient from the structuring element of view and Canny edge detector are widely used. In this paper, we propose the detection algorithm about the people's approach to the subway PSD to prevent the safety accidents by using Canny edge detector and Hough transform and the computer simulation shows the results.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.63 no.5
• /
• pp.676-682
• /
• 2014

In this paper, precipitation echo(PRE) and non-precipitaion echo(N-PRE)(including ground echo and clear echo) through weather radar data are identified with the aid of neuro-fuzzy algorithm. The accuracy of the radar information is lowered because meteorological radar data is mixed with the PRE and N-PRE. So this problem is resolved by using RBFNN and judgement module. Structure expression of weather radar data are analyzed in order to classify PRE and N-PRE. Input variables such as Standard deviation of reflectivity(SDZ), Vertical gradient of reflectivity(VGZ), Spin change(SPN), Frequency(FR), cumulation reflectivity during 1 hour(1hDZ), and cumulation reflectivity during 2 hour(2hDZ) are made by using weather radar data and then each characteristic of input variable is analyzed. Input data is built up from the selected input variables among these input variables, which have a critical effect on the classification between PRE and N-PRE. Echo judgment module is developed to do echo classification between PRE and N-PRE by using testing dataset. Polynomial-based radial basis function neural networks(RBFNNs) are used as neuro-fuzzy algorithm, and the proposed neuro-fuzzy echo pattern classifier is designed by combining RBFNN with echo judgement module. Finally, the results of the proposed classifier are compared with both CZ and DZ, as well as QC data, and analyzed from the view point of output performance.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.2
• /
• pp.47-52
• /
• 2016

The used digital images in the smart device and small displayer has been a downscaled image. In this paper, the detection of the downscaling image forgery is proposed using the feature vector according to the pixel value's gradients. In the proposed algorithm, AR (Autoregressive) coefficients are computed from pixel value's gradients of the image. These coefficients as the feature vectors are used in the learning of a SVM (Support Vector Machine) classification for the downscaling image forgery detector. On the performance of the proposed algorithm, it is excellent at the downscaling 90% image forgery compare to MFR (Median Filter Residual) scheme that had the same 10-Dim. feature vectors and 686-Dim. SPAM (Subtractive Pixel Adjacency Matrix) scheme. In averaging filtering ($3{\times}3$) and median filtering ($3{\times}3$) images, it has a higher detection ratio. Especially, the measured performances of all items in averaging and median filtering ($3{\times}3$), AUC (Area Under Curve) by the sensitivity and 1-specificity is approached to 1. Thus, it is confirmed that the grade evaluation of the proposed algorithm is 'Excellent (A)'.

• Korean Journal of Remote Sensing
• /
• v.19 no.2
• /
• pp.159-169
• /
• 2003

This paper presents a method for automatic 3-d building reconstruction using high resolution aerial imagery. First, by using edge preserving filtering, noise is eliminated and then images are segmented by watershed algorithm, which preserves location of edge pixels. To extract line segments between control points from boundary of each region, we calculate curvature of each pixel on the boundary and then find the control points. Line segment linking is performed according to direction and length of line segments and the location of line segments is adjusted using gradient magnitudes of all pixels of the line segment. Coplanar grouping and pplygonal patch formation are performed per region by selecting 3-d line segments that are matched using epipolar geometry and flight information. The algorithm has been applied to high resolution aerial images and the results show accurate 3D building reconstruction.