• 한국멀티미디어학회논문지
• /
• 제21권7호
• /
• pp.737-744
• /
• 2018

This paper presents a study on the performance of the higher-order moments for speech/music classification. For a successful speech/music classifier, extracting features that allow direct access to the relevant speech or music specific information is crucial. In addition to the conventional variance-based features, we utilize the higher-order moments of features, such as skewness and kurtosis. Moreover, we investigate the subband decomposition parameters in extracting features, which improves classification accuracy. Experiments on two speech/music datasets, which are publicly available, were performed and show that the higher-order moment features can improve classification accuracy when combined with the conventional variance-based features.

• 동의생리병리학회지
• /
• 제25권6호
• /
• pp.1108-1112
• /
• 2011

Voice can be widely used to classify the four constitution types and to recognize one's health condition from extracting meaningful features as physical quantity in traditional Korean medicine or Western medicine. In this paper, we proposed the method to update the standard operating procedure (SOP) to acquire and record voices for extracting stable vocal features since they are sensitive to the variation of a subject's utterance. At first, we obtained pitch frequencies from vowels and the sentence and intensity form the sentence as features with voices acquired under subjects' utterance conditions and then the deviation ratios of features from median values according to the utterance conditions were obtained and the condition to minimize the ratio was selected as a new SOP. As a result, we decided the SOP for a subject to utter vowels with the length of 2s~1s and sentences with over 2s interval between them after practice, in consideration of the deviation and qualitative requirements. Stable voice features obtained from updated SOP produce accurate diagnosis, which will be developed and simplified for using in the u-Healthcare system of personalized medicine.

• 한국멀티미디어학회논문지
• /
• 제23권9호
• /
• pp.1155-1163
• /
• 2020

Deep learning approach is emerging as a new way to improve the accuracy of tree species identification using bark image. However, the approach has not been studied enough because it is confronted with the problem of acquiring a large volume of bark image dataset. This study solved this problem by utilizing a pretrained off-the-shelf DCNN model. It compares the discrimination power of bark features extracted by each DCNN model. Then it extracts the features by using a selected DCNN model and feeds them to a multi-layer perceptron (MLP). We found out that the ResNet50 model is effective in extracting bark features and the MLP could be trained well with the features reduced by the principal component analysis. The proposed approach gives accuracy of 99.1% and 98.4% for BarkTex and Trunk12 datasets respectively.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1996년도 춘계학술대회 논문집
• /
• pp.581-586
• /
• 1996

This paper presents a feature recognition system for recognizing and extracting feature information needed for machining from design data contained in the CAD database of AutoCAD system. The developed system carries out feature recognition from an orthographic view of a press mold containing not only atomic features such as holes, pockets, and slots, but also compound features. Based on the result of feature recognition, it generates a 3-D modeling of the press mold. Especially, The feature recognition part is designed for detecting feature styles according to feature definition and classification, extracting parameters for various atomic features, and constructing necessary data structures for the recognized features.

• Journal of information and communication convergence engineering
• /
• 제1권1호
• /
• pp.53-57
• /
• 2003

Chaos theory is a study researching the irregular, unpredictable behavior of deterministic and non-linear dynamical system. The interpretation using Chaos makes us evaluate characteristic existing in status space of system by tine series, so that the extraction of Chaos characteristic understanding and those characteristics enables us to do high precision interpretation. Therefore, This paper propose the new method which is adopted in extracting character features and recognizing characters using the Chaos Theory. Firstly, it gets features of mesh feature, projection feature and cross distance feature from input character images. And their feature is converted into time series data. Then using the modified Henon system suggested in this paper, it gets last features of character image after calculating Box-counting dimension, Natural Measure, information bit and information dimension which are meant fractal dimension. Finally, character recognition is performed by statistically finding out the each information bit showing the minimum difference against the normalized pattern database. An experimental result shows 99% character classification rates for 2,350 Korean characters (Hangul) using proposed method in this paper.

• Nuclear Engineering and Technology
• /
• 제55권6호
• /
• pp.2246-2255
• /
• 2023

Evacuation time estimation (ETE) is crucial for the effective implementation of resident protection measures as well as planning, owing to its applicability to nuclear emergencies. However, as confirmed in the Fukushima case, the ETE performed by nuclear operators does not reflect behavioral features, exposing thus, gaps that are likely to appear in real-world situations. Existing research methods including surveys and interviews have limitations in extracting highly feasible behavioral features. To overcome these limitations, we propose a VR-based immersive experiment system. The VR system realistically simulates nuclear emergencies by structuring existing disasters and human decision processes in response to the disasters. Evacuation behavioral features were quantitatively extracted through the proposed experiment system, and this system was systematically verified by statistical analysis and a comparative study of experimental results based on previous research. In addition, as part of future work, an application method that can simulate multi-level evacuation dynamics was proposed. The proposed experiment system is significant in presenting an innovative methodology for quantitatively extracting human behavioral features that have not been comprehensively studied in evacuation. It is expected that more realistic evacuation behavioral features can be collected through additional experiments and studies of various evacuation factors in the future.

• Journal of Electrical Engineering and information Science
• /
• 제1권2호
• /
• pp.146-146
• /
• 1996

Although knowledge discovery is increasingly important in databases, the discovered knowledge sets may not be effectively used for application domains. It is partly because knowledge discovery does not take user's interests into account, and too many knowledge sets are discovered to handle efficiently. We believe that user's interests are conveyed by a query and if a nested query is concerned it may include a user's thought process. This paper describes a novel concept for discovering knowledge sets based on query processing. Knowledge discovery process is performed by: extracting features from databases, spanning features to generate range features, and constituting a knowledge set. The contributions of this paper include the following: (1) not only simple queries but also nested queries are considered to discover knowledge sets regarding user's interests and user's thought process, (2) not only positive examples (answer to a query) but also negative examples are considered to discover knowledge sets regarding database abstraction and database exceptions, and (3) finally, the discovered knowledge sets are quantified.

• Journal of Electrical Engineering and information Science
• /
• 제1권2호
• /
• pp.145-156
• /
• 1996

Although knowledge discovery is increasingly important in databases, the discovered knowledge sets may not be effectively used for application domains. It is partly because knowledge discovery does not take user's interests into account, and too many knowledge sets are discovered to handle efficiently. We believe that user's interests are conveyed by a query and if a nested query is concerned it may include a user's thought process. This paper describes a novel concept for discovering knowledge sets based on query processing. Knowledge discovery process is performed by: extracting features from databases, spanning features to generate range features, and constituting a knowledge set. The contributions of this paper include the following: (1) not only simple queries but also nested queries are considered to discover knowledge sets regarding user's interests and user's thought process, (2) not only positive examples (answer to a query) but also negative examples are considered to discover knowledge sets regarding database abstraction and database exceptions, and (3) finally, the discovered knowledge sets are quantified.

• 한국컴퓨터정보학회논문지
• /
• 제11권1호
• /
• pp.139-145
• /
• 2006

디지털 영상에서의 특징점 추출 기술은 로봇비전, 의료영상 진단시스템 및 비디오 전송과 같은 분야 등에서 많이 응용되고 있다. 디지털 영상에서 특징점을 추출하는 방법에는 비선형 그래디언트, 비선형 라프라시안, 엔트로피와 같은 필터들이 있다. 그런데 인간의 시각에서 영상의 특징이 형성되는 과정을 살펴보면, 밝은 영역보다는 어두운 영역에서의 특징에 더 민감한 특성을 가지고 있으므로 기존의 필터로써 특징점을 추출하는데 효과적이지 못하다. 본 논문에서는 국부영역의 밝기를 고려하는 특징점 추출 필터들을 제안한다. 이들 필터들은 연산이 간단하여 매우 신속하게 특징점을 추출할 수 있으며, 국부적인 밝기를 고려하지만 기존의 엔트로피 연산자가 지닌 단점을 극복하여 어두운 영역에서의 미세한 밝기 변화에는 강건한 특성을 가지는 특성을 지닌다. 실험결과 다양한 밝기변화와 국부영역에 걸쳐 매우 뛰어난 특징점 추출결과를 나타내었다.

• 한국컴퓨터정보학회논문지
• /
• 제6권1호
• /
• pp.31-37
• /
• 2001

인간의 시각현상은 밝은 영역보다 어두운 영역의 특징점 추출에 더 민감하다. 그러므로 특징점 추출 연산자는 인간의 시각체계와 유사하게 물체를 인식하기 위해 국부적인 밝기를 고려해야 한다. 일반적으로 지금까지의 특징점 추출 연산자는 계산량이 많거나 다변수를 이용해야 하는 문제점을 가지고 있다. 본 논문에서는 이러한 단점을 극복하는 새로운연산자가 제안된다. 이 연산자는 실현이 매우 간단할 뿐만 아니라 합성영상과 실영상에 적용한 결과 우수한 성능을 나타내었다.