• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1999년도 추계종합학술대회 논문집
• /
• pp.672-675
• /
• 1999

In this paper, we propose a MLP neural network architecture and feature extraction for Korean syllable recognition. In the proposed syllable recognition system, firstly onset is classified by onset classification neural network. And the results information of onset classification neural network are used for feature selection of imput patterns vector. The feature extraction of Korean syllables is based on sonority. Using the threshold rate separate the syllable. The results of separation are used for feature of onset. nucleus and coda. ETRI's SAMDORI has been used by speech DB. The recognition rate is 96% in the speaker dependent and 93.3% in the speaker independent.

• 말소리와 음성과학
• /
• 제9권4호
• /
• pp.91-97
• /
• 2017

This study categorizes several voice characteristics by subjective listening assessment, and investigates correlation between voice characteristics and speech feature parameters. A model was developed to classify voice characteristics into the defined categories using SVM algorithm. To do this, we extracted various speech feature parameters from speech database for men in their 20s, and derived statistically significant parameters correlated with voice characteristics through ANOVA analysis. Then, these derived parameters were applied to the proposed SVM model. The experimental results showed that it is possible to obtain some speech feature parameters significantly correlated with the voice characteristics, and that the proposed model achieves the classification accuracies of 88.5% on average.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1997년도 추계학술대회 논문집 학회본부
• /
• pp.671-674
• /
• 1997

The objective of this paper is to present a new feature-vector extraction method for the automatic detection and classification of power quality(PQ) disturbances, where FIT, DWT(Discrete Wavelet Transform), and Fisher's criterion are utilized to extract an appropriate feature vector. In particular, the proposed classifier consists of three parts: i.e., (i) automatic detection of PQ disturbances, where the wavelet transform and signal power estimation method are utilized to detect each disturbance, (ii) feature vector extraction from the detected disturbance, and (iii) automatic classification, where Multi-Layer Perceptron(MLP) is used to classify each disturbance from the corresponding extracted feature vector. To demonstrate the performance and applicability of the proposed classification algorithm, some test results obtained by analyzing 10-class power quality disturbances are also provided.

• Tribology and Lubricants
• /
• 제15권2호
• /
• pp.206-211
• /
• 1999

Wear debris morphology is closely related to the wear mode and mechanism occured. Image recognition of wear debris is, therefore, a powerful tool in wear monitoring. But it has usually required expert's experience and the results could be too subjective. Development of automatic tools for wear debris recognition is needed to solve this problem. In this work, an algorithm for automatic wear debris recognition was suggested and implemented by PC base software. The presented method defined a characteristic 3-dimensional feature space where typical types of wear debris were separately located by the knowledge-based system and compared the similarity of object wear debris concerned. The 3-dimensional feature space was obtained from multiple feature vectors by using a multi-dimensional scaling technique. The results showed that the presented automatic wear debris recognition was satisfactory in many cases application.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2002년도 추계학술대회 논문집
• /
• pp.353-358
• /
• 2002

A feature-based inspection planning system is proposed in this research to develop more efficient measuring methodology for the OMM (On-machine measurement) for complicated workpiece having many primitive form features. This paper focuses on the development of the CAIP (computer-aided inspection system) methodologies. The optimum inspection sequences for the features are determined by analyzing the feature information such as the nested relations and the possible probe approaching directions of the features, and forming feature groups. A series of heuristic rules are developed to accomplish it. Also, each feature is decomposed into its constituent geometric elements, and then the number of sampling points, the locations of the measuring point, the optimum probing path are determined by applying the fuzzy logic, Hammersley's method, and the TSP algorithm. To verify the proposed methodologies, simulations are carried out and the results are analyzed.

• 한국시뮬레이션학회논문지
• /
• 제14권3호
• /
• pp.137-146
• /
• 2005

In parametric approach to a signature verification, it generally uses so many redundant features unsuitable for each individual signature that it causes harm, instead. This paper proposes a method of determining personalized weights of a feature set in signature verification with parametric approach by identifying the characteristics of each feature. For an individual signature, we define a degree of how difficult it is for any other person to forge the one's (called 'DFD' as the Degree of Forgery Difficulty). According to the statistical characteristics and the intuitional characteristics of each feature, the standard features are classified into four types. Four types of DFD functions are defined and applied into the distance calculation as a personalized weight factor. Using this method, the error rate of signature verification is reduced and the variation of the performance is less sensitive to the changes of decision threshold.

• 지능정보연구
• /
• 제2권1호
• /
• pp.45-58
• /
• 1996

This paper describes an a, pp.oach to recognizing composite features of prismatic parts. AAG (Attribute Adjacency Graph) is adopted as the basis of describing basic feature, but it is extended to enhance the expressive power of AAG by adding face type, angles between faces and normal vectors. Our a, pp.oach is called Extended AAG (EAAG). To simplify the recognition procedure, feature classification tree is built using the graph types of EEA and the number of EAD's. Algorithms to find open faces and dimensions of features are exemplified and used in decomposing composite feature. The processing sequence of recognized features is automatically determined during the decomposition process of composite features.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2001년도 춘계학술발표논문집 (상)
• /
• pp.31-34
• /
• 2001

지리 정보 자원을 서로 공유하기 위한 표준으로 OGC의 OpenGIS simple feature specification이 있다. Feature의 집합을 다루는 Featurelterator 인터페이스에서 기존의 인터페이스는 서버에 새로운 Feature나 Geometry 객체를 생성하고 클라이언트가 이를 다시 접근 해야하는 비효율성이 존재한다. 그래서 서버에 접근하는 비용을 줄이기 위해, 클라이언트가 서버의 Feature들을 접근하는 방법인, 새로운 get_WKSGeometries 인터페이스를 제안하고 구현한다. 이 인터페이스는 서버에 객체를 생성하는 것이 아니라, WKS 형태로 Feature의 데이터를 클라이언트에 직접 전달함으로써 클라이언트가 서버에 재접근을 요구하지 않는다. 따라서, 이 논문에서 제안한 get_WKSGeometries 인터페이스를 이용하면, 클라이언트가 서버에 접근하는 비용을 줄이고 성능을 향상시키도록 하였다. 아울러 구현 명세를 구현하고 새로운 인터페이스를 제안 검증하였다.

• 한국멀티미디어학회논문지
• /
• 제21권12호
• /
• pp.1473-1480
• /
• 2018

The extraction and recognition of human motion characteristics need to combine biometrics to determine and judge human behavior in the movement and distinguish individual identities. The so-called biometric technology, the specific operation is the use of the body's inherent biological characteristics of individual identity authentication, the most noteworthy feature is the invariance and uniqueness. In the past, the behavior recognition technology based on the single characteristic was too restrictive, in this paper, we proposed a mixed feature which combined global silhouette feature and local optical flow feature, and this combined representation was used for human action recognition. And we will use the KTH database to train and test the recognition system. Experiments have been very desirable results.

• 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
• /
• 한국정보과학회언어공학연구회 2017년도 제29회 한글 및 한국어 정보처리 학술대회
• /
• pp.298-300
• /
• 2017

본 논문에서는 단순하지만 효과적인 단어 표현 방법인 Bags of Features에 대한 비교 실험을 수행한다. Bags of Features는 어휘집의 크기에 제한이 없으며, 문자 단위의 정보를 반영하고, 벡터화 과정에서 신경망 구조에 의존하지 않는 단어 표현 방법이다. 영어 품사 부착 실험을 사용하여 실험한 결과, one-hot 인코딩을 사용한 모델과 대비하여 학습 데이터에 존재하지 않는 단어의 경우 49.68%, 전체 부착 정확도는 0.96% 향상이 관찰되었다. 또한, Bags of Features를 사용한 모델은 기존의 영어 품사 부착 분야의 최첨단 모델들 중 학습 데이터 외의 추가적인 데이터를 활용하지 않는 모델들과 비견할 만한 성능을 보였다.