• 전자공학회논문지B
• /
• 제33B권5호
• /
• pp.102-108
• /
• 1996

Adaptive resonance theory (ART2) characterized by its built-in mechanism of handling the stability-plasticity switching and by the adaptive learning without forgetting informations learned in the past, is based on an unsupervised template matching. We propose an improved tow-stage learning algorithm for aRT2: the original unsupervised learning followed by a new supervised learning. Each of the output nodes, after the unsupervised learning, is labeled according to the category informations to reinforce the template pattern associated with the target output node belonging to the same category some dominant classes from exhausting a finite number of template patterns in ART2 inefficiently. Experimental results on a set of 2545 FLIR images show that the ART2 trained by the two-stage learning algorithm yields better accuracy than the original ART2, regardless of th esize of the network and the methods of evaluating the accuracy. This improvement shows the effectiveness of the two-stage learning process.

• Journal of Information Processing Systems
• /
• 제9권2호
• /
• pp.301-314
• /
• 2013

In a parallel processing system, Multi-stage Interconnection Networks (MINs) play a vital role in making the network reliable and cost effective. The MIN is an important piece of architecture for a multiprocessor system, and it has a good impact in the field of communication. Optical Multi-stage Interconnection Networks (OMINs) are the advanced version of MINs. The main problem with OMINs is crosstalk. This paper, presents the (1) Destination Based Modified Omega Network (DBMON) and the (2) Destination Based Scheduling Algorithm (DBSA). DBSA does the scheduling for a source and their corresponding destination address for messages transmission and these scheduled addresses are passed through DBMON. Furthermore, the performance of DBMON is compared with the Crosstalk-Free Modified Omega Network (CFMON). CFMON also minimizes the crosstalk in a minimum number of passes. Results show that DBMON is better than CFMON in terms of the average number of passes and execution time. DBSA can transmit all the messages in only two passes from any source to any destination, through DBMON and without crosstalk. This network is the modified form of the original omega network. Crosstalk minimization is the main objective of the proposed algorithm and proposed network.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1997년도 한국자동제어학술회의논문집; 한국전력공사 서울연수원; 17-18 Oct. 1997
• /
• pp.1758-1761
• /
• 1997

The problem of long range capacity expansion planing for chemical processing network under uncertain demand forecast secnarios is addressed. This optimization problem involves capactiy expansion timing and sizing of each chemical processing unit to maximize the expected net present value considering the deviation of net present values and the excess capacity over a given time horizon. A multiperiod mixed integer nonlinear programming optimization model that is both solution and modle robust for any realization of demand scenarios is developed using the two-stage stochastic programming algorithm. Two example problems are considered to illustrate the effectiveness of the model.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권4호
• /
• pp.1479-1501
• /
• 2020

Dependable functioning of dynamic networks is essential for delivering ubiquitous services. Faults are the root causes of network outages. The comparison diagnosis model, which automates fault's identification, is one of the leading approaches to attain network dependability. Most of the existing research has focused on stationary networks. Nonetheless, the time-free comparison model imposes no time constraints on the system under considerations, and it suits most of the diagnosis requirements of dynamic networks. This paper presents a novel protocol that diagnoses faulty nodes in diagnosable dynamic networks. The proposed protocol comprises two stages, a testing stage, which uses the time-free comparison model to diagnose faulty neighbour nodes, and a disseminating stage, which leverages a Random Linear Network Coding (RLNC) technique to disseminate the partial view of nodes. We analysed and evaluated the performance of the proposed protocol under various scenarios, considering two metrics: communication overhead and diagnosis time. The simulation results revealed that the proposed protocol diagnoses different types of faults in dynamic networks. Compared with most related protocols, our proposed protocol has very low communication overhead and diagnosis time. These results demonstrated that the proposed protocol is energy-efficient, scalable, and robust.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1990년도 한국자동제어학술회의논문집(국제학술편); KOEX, Seoul; 26-27 Oct. 1990
• /
• pp.1259-1264
• /
• 1990

During the past two decades, a lot of researches have been done on the synthesis of grassroot heat exchanger networks(HEN). However, few have been dedicated to retrofit of existing heat exchanger networks, which usually use more amount of utilities (i.e. steam and/or cooling water) than the minimum requirements. This excess gives motivation of trades-off between energy saving and rearranging investment. In this paper, an algorithmic-evolutionary synthesis procedure for retrofitting heat exchanger networks is proposed. It consists of two stages. First, after the amount of maximum energy recovery(MER) is computed, a grass-root network featuring minimum number of units(MNU) is synthesized. In this stage, a systematic procedure of synthesizing MNU networks is presented. It is based upon the concept of pinch, from which networks are synthesized in a logical way by the heuristics verified by the pinch technology. In the second stage, since an initial feasible network is synthesized based on the pre-analysis result of MER and must-matches, an assignment problem between new and existing units is solved to minimize total required additional areas. After the existing units are assigned, the network can be improved by switching some units. For this purpose, an improvement problem is formulated and solved to utilize the areas of existing units as much as possible. An example is used to demonstrate the effectiveness of the proposed method.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 하계학술대회 논문집 G
• /
• pp.2923-2925
• /
• 1999

The objective of camera calibration is to determine the internal optical characteristics of camera and 3D position and orientation of camera with respect to the real world. Calibration procedure applicable to general purpose cameras and lenses. The general method to revise the accuracy rate of calibration is using mathematical distortion of lens. The effective og calibration show big difference in proportion to distortion of camera lens. In this paper, we propose the method which calibration distortion model by using neural network. The neural network model implicity contains all the distortion model. We can predict the high accuracy of calibration method proposed in this paper. Neural network can set properly the distortion model which has difficulty to estimate exactly in general method. The performance of the proposed neural network approach is compared with the well-known Tsai's two stage method in terms of calibration errors. The results show that the proposed approach gives much more stable and acceptabke calibration error over Tsai's two stage method regardless of camera resolution and camera angle.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권8호
• /
• pp.3873-3887
• /
• 2018

In this paper, we propose a novel method to recover images corrupted by impulse noise. The proposed method uses two stages: noise detection and filtering. In the first stage, we use pixel values, rank-ordered logarithmic difference values, and median values to train a neural-network-based impulse noise detector. After training, we apply the network to detect noisy pixels in images. In the next stage, we use group-based weighted couple sparse representation to filter the noisy pixels. During this second stage, conventional methods generally use only clean pixels to recover corrupted pixels, which can yield unsuccessful dictionary learning if the noise density is high and the number of useful clean pixels is inadequate. Therefore, we use reconstructed pixels to balance the deficiency. Experimental results show that the proposed noise detector has better performance than the conventional noise detectors. Also, with the information of noisy pixel location, the proposed impulse-noise removal method performs better than the conventional methods, through the recovered images resulting in better quality.

• Asia pacific journal of information systems
• /
• 제11권4호
• /
• pp.99-111
• /
• 2001

The prediction of stock price index is a very difficult problem because of the complexity of stock market data. It has been studied by a number of researchers since they strongly affect other economic and financial parameters. The movement of stock price index has a series of change points due to the strategies of institutional investors. This study presents a two-stage forecasting model of stock price index using change-point detection and artificial neural networks. The basic concept of this proposed model is to obtain intervals divided by change points, to identify them as change-point groups, and to use them in stock price index forecasting. First, the proposed model tries to detect successive change points in stock price index. Then, the model forecasts the change-point group with the backpropagation neural network(BPN). Finally, the model forecasts the output with BPN. This study then examines the predictability of the integrated neural network model for stock price index forecasting using change-point detection.

• ETRI Journal
• /
• 제40권6호
• /
• pp.774-787
• /
• 2018

Arabic handwriting segmentation and recognition is an area of research that has not yet been fully understood. Dealing with Arabic ligature segmentation, where the Arabic characters are connected and unconstrained naturally, is one of the fundamental problems when dealing with the Arabic script. Arabic character-recognition techniques consider ligatures as new classes in addition to the classes of the Arabic characters. This paper introduces an enhanced technique for Arabic handwriting recognition using the deep belief network (DBN) and a new morphological algorithm for ligature segmentation. There are two main stages for the implementation of this technique. The first stage involves an enhanced technique of the Sari segmentation algorithm, where a new ligature segmentation algorithm is developed. The second stage involves the Arabic character recognition using DBNs and support vector machines (SVMs). The two stages are tested on the IFN/ENIT and HACDB databases, and the results obtained proved the effectiveness of the proposed algorithm compared with other existing systems.

• 제어로봇시스템학회논문지
• /
• 제12권3호
• /
• pp.241-247
• /
• 2006

We propose a character recognition system to extract the component reference names from printed circuit boards (PCBs) automatically. The names are written in horizontal, vertical, reverse-horizontal and reverse-vertical directions. Also various symbols and figures are included in PCBs. To recognize the character and orientation effectively, we divide the recognizer into two stages: character classification stage and orientation classification stage. The character classification stage consists of two sub-recognizers and a verifier. The rotaion-invarint features of input pattern are then used to identify the character independent of orientation. Each recognizer is implemented as a neural network, and the weight values of verifier are obtained by genetic algorithm. In the orientation classification stage, the input pattern is compared with reference patterns to identify the orientation. Experimental results are presented to verify the usefulness of the proposed system.