• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.407-410
• /
• 2001

최근 영상 검색(retrieval)과 분류(classification)에서 질감 특징(texture feature)을 이용한 연구들이 활발하게 진행되고 있다. 본 논문에서는 효율적인 질감 특징 추출을 위해 명암도 상호발생 행렬법(gray level co-occurrence matrix)과 웨이블릿 변환(wavelet transform)을 이용하여 질감의 특징을 추출한 후 특징의 중요도에 따라서 가중치를 부여하는 방법을 제안한다. 이렇게 추출된 가중치 대표 벡터들을 기반으로 베이시안 분류기(Bayesian classifier)를 통해 임의의 질감을 인식하였다.

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10a
• /
• pp.704-706
• /
• 1991

In CAD/CAM system, many efforts are made to automate the converting process from drawling information to manufacturing Information. The most difficult step In this procedure is utilizing 2 dimensional drawing information In order to formulate 3 dimensional w&ling information. This paper emphasizes to mWe automatically series of convertirg steps which provide 3 dimensional wire frame, surface and solid modeling using feature recognition rules. With the standardization of design process and the recognition rule as preceding steps, it shows a good application tool to interface the design and manufacturing procedures in PC-Level CAD/CAM system.

• Proceedings of the IEEK Conference
• /
• 2003.11a
• /
• pp.553-556
• /
• 2003

In the field of content-based image retrieval, various mathematical low-level features have been proposed to describe the perceptual content of images. Since most of the features are assumed to be independent of each other, one feature is extracted from images without any consideration of the other features. Recently proposed CCE and SCFT taking advantage of the correlation between color and texture have shown relatively good performance. In this paper, the performance of CCE, SCFT, and the traditional regular weighted comparison method are evaluated. Simulation results with natural images have shown that CCE outperforms the other methods.

• ETRI Journal
• /
• v.23 no.2
• /
• pp.61-70
• /
• 2001

In this paper, we propose an efficient method for personal identification by analyzing iris patterns that have a high level of stability and distinctiveness. To improve the efficiency and accuracy of the proposed system, we present a new approach to making a feature vector compact and efficient by using wavelet transform, and two straightforward but efficient mechanisms for a competitive learning method such as a weight vector initialization and the winner selection. With all of these novel mechanisms, the experimental results showed that the proposed system could be used for personal identification in an efficient and effective manner.

• Proceedings of the Korean Society for Quality Management Conference
• /
• 2004.04a
• /
• pp.148-151
• /
• 2004

This study presents a new intelligent yield management methodology which can forecast the yield level of a production unit based on features' behaviors. In this proposed methodology, we identify the existing features using C5.0 that are combination of nodes (i.e., variables) in the decision tree generated by C5.0, use SOM(Self-Organizing Map) neural networks in oder to extract the feature's patterns and classify, and then make features' control rules using C5.0.

• The Transactions of the Korean Institute of Electrical Engineers B
• /
• v.54 no.10
• /
• pp.470-476
• /
• 2005

In this paper PU control scheme is presented. The proposed scheme has following features. The one is oft-starting method which is used for preventing to flow large current in motor phase winding when motor starts. The ther is the selection of the level of the over current. The first feature is implemented by increasing the PWM duty lowly, the second feature is implemented by limiting the magnitude of the phase current level by which the over heat f motor by copper losses and magnetic saturation decreases. By the analysis using FEM considering load condition, the peed of mode transition from PW to single pulse control is selected and confirmed by simulation that there is no ver current occurs during the mode transition. For the verification of proposed scheme, the simulation using MATLAB Simulink with considering non-linearity of inductance profile from FEM analysis is performed and the experiment with SRM drive system which has the DSP controller and single Phase SRM are peformed.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.18 no.5
• /
• pp.757-772
• /
• 1993

This paper describes the knowledge extraction for image understanding in knowledge based system. The current set of low level processes operate on the numerical pixel arrays, to segment the image into region and to convert the image into directional image, and to calculate feature for these regions. The current set of intermedate level processes operate on the results of earlier knowledge source to build more complex representations of the data. We have grouped into thee categories : feature based classification, geometric token relation, perceptual organization and grouping.

• Journal of Information Processing Systems
• /
• v.14 no.1
• /
• pp.218-227
• /
• 2018

Vocal effort detection is important for both robust speech recognition and speaker recognition. In this paper, the spectral information entropy feature which contains more salient information regarding the vocal effort level is firstly proposed. Then, the model fusion method based on complementary model is presented to recognize vocal effort level. Experiments are conducted on isolated words test set, and the results show the spectral information entropy has the best performance among the three kinds of features. Meanwhile, the recognition accuracy of all vocal effort levels reaches 81.6%. Thus, potential of the proposed method is demonstrated.

• Journal of the Korea Society of Computer and Information
• /
• v.13 no.5
• /
• pp.53-67
• /
• 2008

At present, for reuse of similar domains between feature model and class model. researches of transformation at the model level and of transformation using ontology between two models are being made. but consistent transformation through metamodel is not made. And the factors of modeling transformation targets are not sufficient, and especially, automatic transformation algorithm and supporting tools are not provided so reuse of domains between models is not activated. This paper proposes a method of transformation from feature model to class model using ontology on the metamodel. For this, it re-establishes the metamodel of feature model, class model, and ontology, and it defines the properties of modelling factors for each metamodel. Based on the properties, it defines the profiles of transformation rules between feature mndel and ontology, and between ontology and class model, using set theory and propositional calculus. For automation of the transformation, it creates transformation algorithm and supporting tools. Using the proposed transformation rules and tools, real application is made through Electronic Approval System. Through this, it is possible to transform from the existing constructed feature model to the class model and to use it again for a different development method. Especially, it is Possible to remove ambiguity of semantic transformation using ontology, and automation of transformation maintains consistence between models.

• Journal of Computing Science and Engineering
• /
• v.9 no.4
• /
• pp.204-213
• /
• 2015

Under the conditional independence assumption among local features, the Naive Bayes Nearest Neighbor (NBNN) classifier has been recently proposed and performs classification without any training or quantization phases. While the original NBNN shows high classification accuracy without adopting an explicit training phase, the conditional independence among local features is against the compositionality of objects indicating that different, but related parts of an object appear together. As a result, the assumption of the conditional independence weakens the accuracy of classification techniques based on NBNN. In this work, we look into this issue, and propose a novel Bayesian network for an NBNN based classification to consider the conditional dependence among features. To achieve our goal, we extract a high-level feature and its corresponding, multiple low-level features for each image patch. We then represent them based on a simple, two-level layered Bayesian network, and design its classification function considering our Bayesian network. To achieve low memory requirement and fast query-time performance, we further optimize our representation and classification function, named relation-based Bayesian network, by considering and representing the relationship between a high-level feature and its low-level features into a compact relation vector, whose dimensionality is the same as the number of low-level features, e.g., four elements in our tests. We have demonstrated the benefits of our method over the original NBNN and its recent improvement, and local NBNN in two different benchmarks. Our method shows improved accuracy, up to 27% against the tested methods. This high accuracy is mainly due to consideration of the conditional dependences between high-level and its corresponding low-level features.