• Journal of the Korean Institute of Telematics and Electronics
• /
• v.25 no.8
• /
• pp.1010-1018
• /
• 1988

This paper proposes a MOS circuit extractor which extracts a netlist from the hierarchical mask information, for the verification tools. To utilize the regularity and the simple representation of the hierarchical circuit, and to reduce the debug cycle of design, verification, and modification, we propose a hierarvhical and incremental circuit extraction algorithm. In flat circuit extraction stage, the multiple storage quad tree is used as an internal data structure. Incremental circuit extraction using the hierarchical structure is made possible, to reduce the re-extraction time of the modified circuit.

• Journal of KIISE:Databases
• /
• v.34 no.4
• /
• pp.343-356
• /
• 2007

Due to advances in position monitoring technologies such as global positioning systems and sensor networks, recent position information of moving objects has the form of streaming data which are updated continuously and rapidly. In this paper we propose an efficient trajectory maintenance method that stores the streaming position data of moving objects in the limited size of storage space and estimates past positions based on the stored data. For this, we first propose a new concept of incremental extraction of position information. The incremental extraction means that, whenever a new position is added into the system, we incrementally re-compute the new version of past position data maintained in the system using the current version of past position data and the newly added position. Next, based on the incremental extraction, we present an overall framework that stores position information and estimates past positions in the stream environment. We then propose two polynomial-based methods, line-based and curve-based methods, as the method of estimating the past positions on the framework. We also propose three incremental extraction methods: equi-width, slope-based, and recent-emphasis extraction methods. Experimental results show that the proposed incremental extraction provides the relatively high accuracy (error rate is less than 3%) even though we maintain only a little portion (only 0.1%) of past position information. In particular, the curve-based incremental extraction provides very low error rate of 1.5% even storing 0.1% of total position data. These results indicate that our incremental extraction methods provide an efficient framework for storing the position information of moving objects and estimating the past positions in the stream environment.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.1042-1045
• /
• 2003

In this paper, we propose the speaker identification system that uses vowel that has speaker's characteristic. System is divided to speech feature extraction part and speaker identification part. Speech feature extraction part extracts speaker's feature. Voiced speech has the characteristic that divides speakers. For vowel extraction, formants are used in voiced speech through frequency analysis. Vowel-a that different formants is extracted in text. Pitch, formant, intensity, log area ratio, LP coefficients, cepstral coefficients are used by method to draw characteristic. The cpestral coefficients that show the best performance in speaker identification among several methods are used. Speaker identification part distinguishes speaker using Neural Network. 12 order cepstral coefficients are used learning input data. Neural Network's structure is MLP and learning algorithm is BP (Backpropagation). Hidden nodes and output nodes are incremented. The nodes in the incremental learning neural network are interconnected via weighted links and each node in a layer is generally connected to each node in the succeeding layer leaving the output node to provide output for the network. Though the vowel extract and incremental learning, the proposed system uses low learning data and reduces learning time and improves identification rate.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.8 no.7
• /
• pp.1475-1479
• /
• 2004

High dimensional data requires efficient feature extraction techliques. Though PCA(Principal Component Analysis) is a famous feature extraction method it requires huge memory space and computational cost is high. In this paper we use incremental PCA for feature extraction on high dimensional data. Through experiment we show that proposed method is superior to APEX model.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.7
• /
• pp.2496-2511
• /
• 2014

Face recognition is considered a complex biometrics in the field of image processing mainly due to the constraints imposed by variation in the appearance of facial images. These variations in appearance are affected by differences in expressions and/or occlusions (sunglasses, scarf etc.). This paper discusses incremental Kernel Fisher Discriminate Analysis on sub-classes for dealing with partial occlusions and variant expressions. This framework focuses on the division of classes into fixed size sub-classes for effective feature extraction. For this purpose, it modifies the traditional Linear Discriminant Analysis into incremental approach in the kernel space. Experiments are performed on AR, ORL, Yale B and MIT-CBCL face databases. The results show a significant improvement in face recognition.

• Journal of KIISE:Software and Applications
• /
• v.34 no.1
• /
• pp.11-17
• /
• 2007

One of the popular methods used for pattern classification is the MBR (Memory-Based Reasoning) algorithm. Since it simply computes distances between a test pattern and training patterns or hyperplanes stored in memory, and then assigns the class of the nearest training pattern, it cannot explain how the classification result is obtained. In order to overcome this problem, we propose an incremental teaming algorithm based on RPA (Recursive Partition Averaging) to extract IF-THEN rules that describe regularities inherent in training patterns. But rules generated by RPA eventually show an overfitting phenomenon, because they depend too strongly on the details of given training patterns. Also RPA produces more number of rules than necessary, due to over-partitioning of the pattern space. Consequently, we present the IREA (Incremental Rule Extraction Algorithm) that overcomes overfitting problem by removing useless conditions from rules and reduces the number of rules at the same time. We verify the performance of proposed algorithm using benchmark data sets from UCI Machine Learning Repository.

• Journal of Internet Computing and Services
• /
• v.4 no.3
• /
• pp.39-46
• /
• 2003

Generally, the existing watermarking techniques for copyright protection of a digital image are fragile in geometric distortion and all watermark extractions need the same time regardless of degree of distortion. In this paper, we propose the incremental watermarking technique that used a Complex Wavelet Transform(CWT) in order to solve these problems. The proposed incremental watermarking technique embeds a watermark in a phase component after a CWT with an original image, and a watermark is extracted from an watermarked image by stages. A watermark owner can insist on copyright of an image after comparing a correlation between the extracted watermark and the original watermark if it is larger than the threshold. Also, the incremental watermark extraction determines the extraction time of a watermark by the level of distortion. The proposed technique through performance evaluation displayed that it was robust in geometric distortion than the existing watermarking technique.

• Korean Management Science Review
• /
• v.20 no.2
• /
• pp.1-16
• /
• 2003

Previous research on data extraction and integration for data warehousing has concentrated mainly on the relational DBMS or partly on the object-oriented DBMS. Mostly, it describes issues related with the change data (deltas) capture and the incremental update by using the triggering technique of active database systems. But, little attention has been paid to data extraction approaches from other types of source systems like hierarchical DBMS, etc. and from source systems without triggering capability. This paper argues, from the practical point of view, that we need to consider not only the types of information sources and capabilities of ETT tools but also other factors of source systems such as operational characteristics (i.e., whether they support DBMS log, user log or no log, timestamp), and DBMS characteristics (i.e., whether they have the triggering capability or not, etc), in order to find out appropriate data extraction techniques that could be applied to different source systems. Having applied several different data extraction techniques (e.g., DBMS log, user log, triggering, timestamp-based extraction, file comparison) to S bank's source systems (e.g., IMS, DB2, ORACLE, and SAM file), we discovered that data extraction techniques available in a commercial ETT tool do not completely support data extraction from the DBMS log of IMS system. For such IMS systems, a new date extraction technique is proposed which first creates Index database and then updates the data warehouse using the Index database. We illustrates this technique using an example application.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.26 no.1
• /
• pp.56-63
• /
• 2016

In this study, we introduce a design of Fuzzy RBFNNs-based digit recognition system using the incremental-PCA in order to recognize the handwritten digits. The Principal Component Analysis (PCA) is a widely-adopted dimensional reduction algorithm, but it needs high computing overhead for feature extraction in case of using high dimensional images or a large amount of training data. To alleviate such problem, the incremental-PCA is proposed for the computationally efficient processing as well as the incremental learning of high dimensional data in the feature extraction stage. The architecture of Fuzzy Radial Basis Function Neural Networks (RBFNN) consists of three functional modules such as condition, conclusion, and inference part. In the condition part, the input space is partitioned with the use of fuzzy clustering realized by means of the Fuzzy C-Means (FCM) algorithm. Also, it is used instead of gaussian function to consider the characteristic of input data. In the conclusion part, connection weights are used as the extended diverse types in polynomial expression such as constant, linear, quadratic and modified quadratic. Experimental results conducted on the benchmarking MNIST handwritten digit database demonstrate the effectiveness and efficiency of the proposed digit recognition system when compared with other studies.

• Journal of Soil and Groundwater Environment
• /
• v.16 no.6
• /
• pp.1-9
• /
• 2011

A bibliographical review of Gy sampling theory for particulate materials was conducted to provide readers with useful means to reduce errors in soil contamination investigation. According to the Gy theory, the errors caused by the heterogeneous nature of soil include; the fundamental error (FE) caused by physical and chemical constitutional heterogeneity, the grouping and segregation error (GE) aroused from gravitational force, long-range heterogeneous fluctuation error ($CE_2$), the periodic heterogeneity fluctuation error ($CE_3$), and the materialization error (ME) generated during physical process of sample treatment. However, the accurate estimation of $CE_2$ and $CE_3$ cannot be estimated easily and only increasing sampling locations can reduce the magnitude of the errors. In addition, incremental sampling is the only method to reduce GE while grab sampling should be avoided as it introduces uncertainty and errors to the sampling process. Correct preparation and operation of sampling tools are important factors in reducing the incremental delimitation error (DE) and extraction error (EE) which are resulted from physical processes in the sampling. Therefore, Gy sampling theory can be used efficiently in planning a strategy for soil investigations of non-volatile and non-reactive samples.