• IEMEK Journal of Embedded Systems and Applications
• /
• v.17 no.6
• /
• pp.309-317
• /
• 2022

In this paper, we propose a machine learning-based copy-move forgery detection network with dual branches. Because the rotation or scaling operation is frequently involved in copy-move forger, the conventional convolutional neural network is not effectively applied in detecting copy-move tampering. Therefore, we divide the input into rotation-invariant and scaling-invariant features based on the wavelet coefficients. Each of the features is input to different branches having the same structure, and is fused in the combination module. Each branch comprises feature extraction, correlation, and mask decoder modules. In the proposed network, VGG16 is used for the feature extraction module. To check similarity of features generated by the feature extraction module, the conventional correlation module used. Finally, the mask decoder model is applied to develop a pixel-level localization map. We perform experiments on test dataset and compare the proposed method with state-of-the-art tampering localization methods. The results demonstrate that the proposed scheme outperforms the existing approaches.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.2
• /
• pp.237-244
• /
• 2020

This paper describes an implement of Simultaneous Localization and Mapping(SLAM) in two dimensional space. The method uses Invariant Extended Kalman Filter(IEKF), which transforms the state variables and measurement variables so that the transformed variables constitute a linear space when variables called the invariant quantities are kept constant. Therefore, the IEKF guarantees convergence provided in the invariant quantities are kept constant. The proposed IEKF approach uses Lie group matrix for the transformation. The method is tested through simulation, and the results show that the Kalman gain is constant as it is the case for the linear Kalman filter. The coherence between the estimated locations of the vehicle and the detected objects verifies the estimation performance of the method.

• Bulletin of the Korean Mathematical Society
• /
• v.55 no.4
• /
• pp.1221-1230
• /
• 2018

We derive a sharp decomposition formula for the state polytope of the Hilbert point and the Hilbert-Mumford index of reducible varieties by using the decomposition of characters and basic convex geometry. This proof captures the essence of the decomposition of the state polytopes in general, and considerably simplifies an earlier proof by the authors which uses a careful analysis of initial ideals of reducible varieties.

• Bulletin of the Korean Mathematical Society
• /
• v.35 no.4
• /
• pp.689-700
• /
• 1998

We consider an infinite dimensional dynamical system what is called Lattice Dynamical System given by a discrete nonlinear diffusion equation. By assuming the nonlinearity to be a general nonlinear function with mild restrictions, we show that as the diffusion parameter changes the stationery state of the given system undergoes bifurcations from the zero state to a bounded invariant set or a 3- or 4-periodic state in the global phase space of the given system according to the values of the coefficients of the linear part of the given nonlinearity.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.8
• /
• pp.1434-1439
• /
• 2008

Numerical methods for solving the state feedback control problem of linear time invariant system are presented in this paper. The methods are based on Haar wavelet approximation. The properties of Haar wavelet are first presented. The operational matrix of integration and its inverse matrix are then utilized to reduce the state feedback control problem to the solution of algebraic matrix equations. The proposed methods reduce the computation time remarkably. Finally a numerical example is illustrated to demonstrate the validity and applicability of the proposed methods.

• Proceedings of The Korean Society of Agricultural and Forest Meteorology Conference
• /
• 2003.09a
• /
• pp.50-53
• /
• 2003

Models have been developed to estimate leaf wetness duration (LWD) using conventional weather observations, e.g., air temperature, water vapor pressure, and wind speed, which are relatively invariant over space (Pedro and Gillespie, 1982; Gleason et al., 1994; Francl and Panigrahi, 1997).(omitted)

• Journal of IKEEE
• /
• v.23 no.2
• /
• pp.448-453
• /
• 2019

Under the conflicting objectives of maximum user satisfaction and fast launching, there exist great needs for automated mobile app testing. In automated app testing, detection of usage-state changes is one of the most important issues for minimizing human intervention and testing of various usage scenarios. Because conventional approaches utilizing pre-collected training examples can not handle the rapid evolution of apps, we propose a novel method detecting changes in usage-state through graph-entropy. In the proposed method, widgets in a screen shot are recognized through DNNs and 'onverted graphs. We compared the performance of the proposed method with a SIFT (Scale-Invariant Feature Transform) based method on 20 real-world apps. In most cases, our method achieved superior results, but we found some situations where further improvements are required.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1995.10a
• /
• pp.299-303
• /
• 1995

This paper charcterizes dynamics of cold tandem minns, and constructs it state-space model of which are linear time invariant, using subspace method. Step responses particularly show the influence on mass transfer delay. Input-output data set are obtained form nonlinear differential equations including mass transfer delay and nonlinearity. It is shown that the identified state-apace model well approximates the original systems dynamics.

• 제어로봇시스템학회:학술대회논문집
• /
• 1993.10a
• /
• pp.947-952
• /
• 1993

A simplified pole-placement design method is developed by analysing dynamic characteristics of the switching dynamics. Unlike the design procedure of conventional pole-placement, in the proposed method, overall state-space is directly decomposed into two invariant subspaces by the projection operator which is defined in the equivalent system, and then the closed-loop poles are assigned to each subspace independently. Hence, computations for state-feedback gain matrix are easy and simple.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.980-983
• /
• 1996

A new version of the discrete-time optimal FIR (finite impulse response) filter utilizing only the measurements of finite sliding estimation window is suggested for linear time-invariant state-space models. This filter is called the BLUFIR (best linear unbiased finite impulse response) filter since it provides the BLUE (best linear unbiased estimate) of the state obtained from the measurements of the estimation window. It is shown that the BLUFIR filter has the deadbeat property when there are no noises in the estimation window.