• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.11
• /
• pp.465-472
• /
• 2022

In this paper, the style synthesis network is trained to generate style-synthesized video through the style synthesis through training Stylegan and the video synthesis network for video synthesis. In order to improve the point that the gaze or expression does not transfer stably, 3D face restoration technology is applied to control important features such as the pose, gaze, and expression of the head using 3D face information. In addition, by training the discriminators for the dynamics, mouth shape, image, and gaze of the Head2head network, it is possible to create a stable style synthesis video that maintains more probabilities and consistency. Using the FaceForensic dataset and the MetFace dataset, it was confirmed that the performance was increased by converting one video into another video while maintaining the consistent movement of the target face, and generating natural data through video synthesis using 3D face information from the source video's face.

• Journal of Institute of Control, Robotics and Systems
• /
• v.1 no.2
• /
• pp.119-126
• /
• 1995

The purpose of this study is to develop the technique of energy recovery and energy saving by using the optimization of heat exchanger network synthesis. This article proposes a new method of determining the optimal target of a heat exchanger network synthesis problem of which data feature multiple pinch points. The system separation method we suggest here is to subdivide the original system into independent subsystems with one pinch point. The optimal cost target was evaluated and the original pinch rules at each subsystem were employed. The software developed in this study was applied to the Alko prosess, which is an alcohol production process, for the synthesis of heat exchanger network. It was possible to save about 15% of the total annual cost.

• International Journal of Control, Automation, and Systems
• /
• v.6 no.1
• /
• pp.24-34
• /
• 2008

A novel neural network model, termed the standard neural network model (SNNM), similar to the nominal model in linear robust control theory, is suggested to facilitate the synthesis of controllers for delayed (or non-delayed) nonlinear systems composed of neural networks. The model is composed of a linear dynamic system and a bounded static delayed (or non-delayed) nonlinear operator. Based on the global asymptotic stability analysis of SNNMs, Static state-feedback controller and dynamic output feedback controller are designed for the SNNMs to stabilize the closed-loop systems, respectively. The control design equations are shown to be a set of linear matrix inequalities (LMIs) which can be easily solved by various convex optimization algorithms to determine the control signals. Most neural-network-based nonlinear systems with time delays or without time delays can be transformed into the SNNMs for controller synthesis in a unified way. Two application examples are given where the SNNMs are employed to synthesize the feedback stabilizing controllers for an SISO nonlinear system modeled by the neural network, and for a chaotic neural network, respectively. Through these examples, it is demonstrated that the SNNM not only makes controller synthesis of neural-network-based systems much easier, but also provides a new approach to the synthesis of the controllers for the other type of nonlinear systems.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.2
• /
• pp.216-221
• /
• 2004

This paper describes an optimal synthesis method of binary neural network for pattern recognition. Our objective is to minimize the number of connections and the number of neurons in hidden layer by using a Newly Expanded and Truncated Learning Algorithm (NETLA) for the multilayered neural networks. The synthesis method in NETLA uses the Expanded Sum of Product (ESP) of the boolean expressions and is based on the multilayer perceptron. It has an ability to optimize a given binary neural network in the binary space without any iterative learning as the conventional Error Back Propagation (EBP) algorithm. Furthermore, NETLA can reduce the number of the required neurons in hidden layer and the number of connections. Therefore, this learning algorithm can speed up training for the pattern recognition problems. The superiority of NETLA to other learning algorithms is demonstrated by an practical application to the approximation problem of a circular region.

• International Journal of Aeronautical and Space Sciences
• /
• v.7 no.1
• /
• pp.129-136
• /
• 2006

A new guidance synthesis for anti-ship missiles to control impact angle and impact time is proposed in this paper. The flight vehicle is assumed as a 1st order lag system to consider more practical system. The proposed guidance synthesis enhances the survivability of anti-ship missiles because multiple anti-ship missiles with the proposed synthesis can hit the target simultaneously. The control input to satisfy constraints of zero miss distance and impact angle, and the feedforward bias control input to control impact time constitute the guidance law. The former is from trajectory shaping guidance, the latter is from neural network. And particle swarm optimization method is introduced to furnish reference input and output for learning in neural network. The performance of the proposed synthesis in the accuracy of impact time and angle is validated by numerical examples.

• IE interfaces
• /
• v.24 no.3
• /
• pp.187-195
• /
• 2011

The automatic radiopharmaceutical module consists of several 2-way valves, couple of syringes, gas supply unit, heating(cooling) unit and sensors to control the chemical reagents as well as to help the chemical reaction. In order to control the actuators of radiopharmaceutical module, the process is tabulated using spread sheet as like excel. Unlike the common program, a trivial error is too critical to allowed in the process because the error can lead to leak the radioactive reagent and to cause the synthesis equipment failure during synthesizing. Hence, the synthesis process has been validated using graphic simulation while the operator checks the whole process visually and undergoes trial and error. The verification of the synthesis process takes a long time and has a difficulty in finding the error. This study presents a methodology to verify the process algebraically while the radiopharmaceutical module is converted to the network model. The proposed method is validated using actual synthesis process.

• Korean Journal of Hydrosciences
• /
• v.2
• /
• pp.69-84
• /
• 1991

US NWS/NETWORK is applied for the analysis of the flood of July 11-15, 1981 through the Goan-Indogyo reach of the Han River. For the flood hydrography synthesis of the lateral inflows from the major tributaries into the main reach the Cleak method is employed. NETWORK coupled with the Clark method of hydrography synthesis simulated with a fair accuracy the oberved flood hydrograph at the downstream boundary of the routing reach. The dffect of SCS runoff curve number for fributary flood synthesis is evaluated. The characteristics of the station variations and time variations of the flood discharges in the reach is also analyzed.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.9 no.5
• /
• pp.6-11
• /
• 1972

A general active RC network synthesis procedure which realizes any stable transfer function is described. The network elements are only R's, C's and OA's, and the network configuration are well suited for construction using thin-film RC networks and integrated cil'suit operational amplifiers. Poles and transmission zeros can be adjusted independently to each other and are qu;te insensitive to element variations.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.156.3-156
• /
• 2001

This paper handles synthesis and investigation of a neural network guidance law based on pursuit-evasion games. This work considers two-dimensional pursuit-evasion games solved by using the gradient method. The procedure of developing a guidance law from the game-optimal solutions is deeply examined, and important features of the neural network guidance are investigated. The proposed neural network guidance law takes the range, range rate, line-of-sight rate, and heading error as its input variables. By reconstructing the trajectory, the accuracy of the neural network approximation is verified. Afterwards, robustness of the neural network guidance to the autopilot lag, which results from its feedback structure, is investigated ...

• Proceedings of the KIEE Conference
• /
• 2004.11c
• /
• pp.391-393
• /
• 2004

For performance and stability of a synchronized system, we need an efficient Clock Tree Synthesis(CTS) methodology to design clock distribution networks. In a system-on-a-chip(SOC) design environment, CTS effectively distributes clock signals from clock sources to synchronized points on layout design. In this paper, we suggest the pre-layout analysis of the clock network including gated clock, multiple clock, and test mode CTS optimization. This analysis can help to avoid design failure with potential CTS problems from logic designers and supply layout constraints so as to get an optimal clock distribution network. Our new design flow including pre-layout CTS analysis and structural violation checking also contributes to reduce design time significantly.