• Journal of Korea Game Society
• /
• v.20 no.3
• /
• pp.65-74
• /
• 2020

In the shadow map method, if the shadow map is magnified, the shadow has a jagged silhouette. Herein, we propose a soft shadow method that filters reshaped silhouettes analytically. First, the shadow silhouette is reshaped through sub-texel edge detection, which is based on linear or quadratic curve models. Second, an integral shadow filtering algorithm is used to accurately obtain the average shadow intensity from a definite integral estimation. The implementation demonstrates that our solution can effectively eliminate jagged aliasing and efficiently generate soft shadows.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.40 no.5
• /
• pp.250-258
• /
• 2003

In this paper, an explanation regarding on-line identification of a fuzzy system is presented. The fuzzy system to be identified is assumed to be in the type of singleton consequent parts and be represented by a linear combination of fuzzy basis functions. For on-line identification, squared-cosine membership function is introduced to reduce the number of parameters to be identified and make the system consistent and differentiable. Then the parameters of the fuzzy system are identified on-line by the gradient search method and Extended Kalman Filter. Finally, a computer simulation is peformed to illustrate the validity of the suggested algorithms.

• Journal of the Korean Society of Radiology
• /
• v.2 no.1
• /
• pp.23-29
• /
• 2008

The present study proposed a method that dissolves ultrasonographic images into multiple resolutions using wavelet conversion and a boundary detection filter and improves the quality of ultrasonographic images through boundary detection filtering. In order to reduce noises and strengthen edges, the proposed method adjusted selectivity coefficient by area step by step from a low resolution image obtained from wavelet converted images to a high resolution image and performed edge filtering in consideration of direction. Through this method, we generated a selective low pass filtering effect in areas except edges by decreasing the wavelet coefficient for pixels in spot areas, improved continuity by smoothing edges in the tangential direction, and enhanced contrast by thinning in the normal direction. Through an experiment, we compared the filtering method using a non linear anisotropic expansion model and the filtering method using wavelet contraction structure in single resolution.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.10
• /
• pp.166-174
• /
• 1998

Neural networks, having highly nonlinear dynamics by virtue of the distributed nonlinearities and the learing ability, have the potential for the adaptive prediction of nonstationary signals. This paper describes the nonlinear prediction of these signals in two ways; using a nonlinear module and the cascade combination of nonlinear and linear modules. Fully-connected recurrent neural networks (RNNs) and a conventional tapped-delay-line (TDL) filter are used as the nonlinear and linear modules respectively. The dynamic behavior of the proposed predictors is demonstrated for chaotic time series adn speech signals. For the relative comparison of prediction performance, the proposed predictors are compared with a conventional ARMA linear prediction model. Experimental results show that the neural networks based adaptive predictor ourperforms the traditional linear scheme significantly. We also find that the cascade combination predictor is well suitable for the prediction of the time series which contain large variations of signal amplitude.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.1
• /
• pp.139-147
• /
• 2003

Dynamic neural networks have been applied to diverse fields requiring temporal signal processing such as signal prediction. This paper proposes the hybrid network, composed of locally(LRNN) and globally recurrent neural networks(GRNN), to improve dynamics of multilayered recurrent networks(RNN) and then describes nonlinear adaptive prediction using the proposed network as an adaptive filter. The hybrid network consists of IIR-MLP and Elman RNN as LRNN and GRNN, respectively. The proposed network is evaluated in nonlinear signal prediction and compared with Elman RNN and IIR-MLP networks for the relative comparison of prediction performance. Experimental results show that the hybrid network performs better with respect to convergence speed and accuracy, indicating that the proposed network can be a more effective prediction model than conventional multilayered recurrent networks in nonlinear prediction for nonstationary signals.

• Journal of the Institute of Convergence Signal Processing
• /
• v.7 no.2
• /
• pp.53-59
• /
• 2006

Dynamic neural networks have been applied to diverse fields requiring temporal signal processing such as system identification and signal prediction. This paper proposes the gamma neural network(GAM), which uses gamma memory kernel in the hidden layer of feedforward multilayered network, to improve dynamics of networks and then describes nonlinear adaptive prediction using the proposed network as an adaptive filter. The proposed network is evaluated in nonlinear signal prediction and compared with feedforword(FNN) and recurrent neural networks(RNN) for the relative comparison of prediction performance. Simulation results show that the GAM network performs better with respect to the convergence speed and prediction accuracy, indicating that it can be a more effective prediction model than conventional multilayered networks in nonlinear prediction for nonstationary signals.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.7
• /
• pp.50-61
• /
• 1999

\In this paper, we consider the optimal scheduling of detection and tracking parameters in phased array radars to minimize the radar energy required for track maintenance in a cluttered environment. We develop a mathematical model of target detection induced by a search process in phased array radars. In the mathematical development, we take into account the effect of unwanted measurements that may have originated from clutter or false alarms in the detection process. We use and analytic approximation of the modified Riccati equation of the probabilistic data association (PDA) filter to take into account the effect of clutter interference in tracking. Based on the search process and the tracking models, we formulate the optimal scheduling problem into a nonlinear optimal control problem. We solve a constrained nonlinear optimization problem to obtain the solution of the optimal control problem.

• 전기의세계
• /
• v.39 no.12
• /
• pp.21-32
• /
• 1990

퍼지 논리를 이용한 제어시스템에 관하여 핵심 개념을 중심으로 기술하고자 한다. 요약컨데 이 퍼지제어기의 특징은 1) Parallel(distributed) control 2) logic control 3) linguistic control등이며 퍼지 제어가 효과적일 수 있는 제어대상(plant)로서는 수학적 모델을 적용하기 힘든 시스템으로서 경험적으로 또는 수동적인 방법으로 제어가 잘되고 있는 대상을 들 수 있다. 그 뿐만 아니라 간단한 제어기가 필요한 경우로서 보다 효과적인 제어측 Software를 쓰거나 센서 또는 필터없이 사용가능하고, Inverted Penedulum의 자세 제어처럼 정확성보다는 속도 응답 제어가 요구되는 경우 등에 효과적으로 쓸 수 있는 것으로 알려지고 있다. Fuzzy 제어는 지식 베이스의 규모에서 인공지능형 Expert System보다 Compact하고 선형.비선형 플랜트에 공히 이용될 수 있으며, 설계자는 오퍼레이터와의 접촉을 통해 룰을 구축하므로 사용자가 시스템을 이해하기 쉬운 잇점등이 있기도 한다. 그러나 가장 큰 문제는 구축해 놓은 시스템의 안전성(Stability)를 이론적으로 사전에 검증하기가 어렵고, 같은 제어대상이라 할지라도 추론방법, 소속함수의 형태선택, 룰수 등에 따라 제어성능이 바뀔수 있으나, 무엇이 어떤 영향을 주는지 규명되지 않은점 등 여러가지 연구되어야 할 내용이 많이 있다.

• Proceedings of the KIEE Conference
• /
• 2015.07a
• /
• pp.948-949
• /
• 2015

일반적으로 IPMSM의 전압방정식은 d축과 q축이 90도의 위상차를 가지고 있기 때문에 d-q축 간의 상호 인덕턴스를 고려하지 않는다. 하지만 실제로는 d축의 인덕턴스는 q축 전류에 영향을 받으며, 반대로 q축의 인덕턴스도 d축 전류에 영향을 받는다. 따라서 비선형 모델링을 통해 실제 전동기의 형태에 더 가깝게 묘사 하였다. 또한 일반적인 수학식으로 계산하여 Ldq, Lqd를 구해 LPF 필터를 사용하였고 이산적인 최소자승법을 이용한 Gain값을 통해 과도상태에서 더 적합한 LPF와 최소자승법을 비교하는 논문이다.

• Journal of Internet Computing and Services
• /
• v.23 no.4
• /
• pp.21-33
• /
• 2022

Recently, the recommender system has been widely used in various fields such as movies, music, online shopping, and social media, and in the meantime, the recommender model has been developed from correlation analysis through the Apriori model, which can be said to be the first-generation model in the recommender system field. In 2005, many models have been proposed, including deep learning-based models, which are receiving a lot of attention within the recommender model. The recommender model can be classified into a collaborative filtering method, a content-based method, and a hybrid method that uses these two methods integrally. However, these basic methods are gradually losing their status as methodologies in the field as they fail to adapt to internal and external changing factors such as the rapidly changing user-item interaction and the development of big data. On the other hand, the importance of deep learning methodologies in recommender systems is increasing because of its advantages such as nonlinear transformation, representation learning, sequence modeling, and flexibility. In this paper, among deep learning methodologies, RNN, CNN, and GAN-based models suitable for sequential modeling that can accurately and flexibly analyze user-item interactions are classified, compared, and analyzed.