• 한국인터넷방송통신학회논문지
• /
• 제8권6호
• /
• pp.1-7
• /
• 2008

본 논문은 스케일 스페이스 필터링 기법을 활용하여 홍채영상으로부터 사람마다 고유한 특징을 추출하는 방법을 제시한다. 추출되는 특징은 성능이 우수하고 신뢰도가 높아 고속의 자동 인식 시스템의 제작에 활용 될 수 있다. 제시하는 알고리즘은 우선 홍채영상으로부터 홍채 부분을 분리하고 홍채의 중심과 반경을 산출한 후, 노이즈가 심한 부분을 제거하고 2D 형태의 고유한 특징들을 추출한다. 노이즈에 대한 영향을 최소화 하기 위해 스케일 스페이스 필터링이 사용된다. 성능을 입증하기 위해 18명으로부터 얻은 272개의 홍채영상을 대상으로 실험을 수행하였다. 실험결과는 제시하고 있는 알고리즘이 성능과 신뢰도 측면에서 매우 우수함을 보여준다.

• 한국전자통신학회논문지
• /
• 제5권4호
• /
• pp.363-371
• /
• 2010

특징벡터의 분류를 개선시켜 화자독립 음성인식의 오류율을 줄이려는 노력의 일환으로서, 우리는 MFCC의 추출에 있어서 푸리에 스펙트럼을 기울이는 방법이 미치는 효과를 연구한다. 음성신호에 FIR 필터링을 적용하는 효과의 조사도 병행된다. 제안된 방법은 두 가지 독립적인 방법에 의해 평가된다. 즉, 피셔의 차별함수에 의한 방법과 은닉 마코브 모델 및 퍼지 벡터양자화를 사용한 음성인식 오류율 조사 방법이다. 실험 결과, 적절한 파라미터의 선택에 의해 기존의 방법에 비해 10% 정도 낮은 인식 오류율이 얻어짐을 확인하였다.

• 전기학회논문지
• /
• 제67권10호
• /
• pp.1358-1363
• /
• 2018

As the usage of the Internet grows and digital media become more diversified, it has become much easier for digital contents to be distributed and shared. This makes easier to access the desired digital contents. On the other hand, there is an increasing need to protect the copyright of digital works. There are some prevalent ways to protect ownership, but they accompany several disadvantages. Among those ways, watermarking methods have the advantage of ensuring invisibility, but they also have a disadvantage that they are vulnerable to external attacks such as a noise and signal processing. In this paper, we propose the detecting method of illegal contents that is robust against external attacks to protect digital works. We extract HSV and LBP features from images and use Euclidian-based hashing techniques to shorten the searching time on high-dimensional and near-duplicate videos. According to the results, the proposed method showed higher detection rates than that of the Watermarking techniques in terms of the images with fabrications or deformations.

• International Journal of Contents
• /
• 제7권1호
• /
• pp.14-22
• /
• 2011

Temporal Sign Language Recognition (TSLR) from hand motion is an active area of gesture recognition research in facilitating efficient communication with deaf people. TSLR systems consist of two stages: a motion sensing step which extracts useful features from signers' motion and a classification process which classifies these features as a performed sign. This work focuses on two of the research problems, namely unknown time varying signal of sign languages in feature extraction stage and computing complexity and time consumption in classification stage due to a very large sign sequences database. In this paper, we propose a combination of Dynamic Time Warping (DTW) and application of the Single hidden Layer Feedforward Neural networks (SLFNs) trained by Extreme Learning Machine (ELM) to cope the limitations. DTW has several advantages over other approaches in that it can align the length of the time series data to a same prior size, while ELM is a useful technique for classifying these warped features. Our experiment demonstrates the efficiency of the proposed method with the recognition accuracy up to 98.67%. The proposed approach can be generalized to more detailed measurements so as to recognize hand gestures, body motion and facial expression.

• Smart Structures and Systems
• /
• 제22권1호
• /
• pp.1-11
• /
• 2018

This study presents early crack detection of steel fiber-reinforced concrete (SFRC) under varying temperature and humidity conditions using an instantaneous electrical impedance acquisition system. SFRC has the self-sensing capability of electrical impedance without sensor installation thanks to the conductivity of embedded steel fibers, making it possible to effectively monitor cracks initiated in SFRC. However, the electrical impedance is often sensitively changed by environmental effects such as temperature and humidity variations. Thus, the extraction of only crack-induced feature from the measured impedance responses is a crucial issue for the purpose of structural health monitoring. In this study, the instantaneous electrical impedance acquisition system incorporated with SFRC is developed. Then, temperature, humidity and crack initiation effects on the impedance responses are experimentally investigated. Based on the impedance signal pattern observation, it is turned out that the temperature effect is more predominant than the crack initiation and humidity effects. Various crack steps are generated through bending tests, and the corresponding impedance damage indices are extracted by compensating the dominant temperature effect. The test results reveal that propagated cracks as well as early cracks are successfully detected under temperature and humidity variations.

• 한국전기전자재료학회논문지
• /
• 제19권2호
• /
• pp.162-168
• /
• 2006

Recently, diagnosis techniques have been investigated to detect a Partial Discharge associated with a dielectric material defect in a high voltage electrical apparatus, However, the properties of detection technique of Partial Discharge aren't completely understood because the physical process of Partial Discharge. Therefore, this paper analyzes the process on surface discharge of polymer insulator using wavelet transform. Wavelet transform provides a direct quantitative measure of spectral content in the time~frequency domain. As it is important to develop a non-contact method for detecting the kaolin contamination degree, this research analyzes the electromagnetic waves emitted from Partial Discharge using wavelet transform. This result experimentally shows the process of Partial Discharge as a two-dimensional distribution in the time-frequency domain. Feature extraction parameter namely, maximum and average of wavelet coefficients values, wavelet coefficients value at the point of $95\%$ in a histogram and number of maximum wavelet coefficient have used electromagnetic wave signals as input signals in the preprocessing process of neural networks in order to identify kaolin contamination rates. As result, root sum square error was produced by the test with a learning of neural networks obtained 0.00828.

• 한국전기전자재료학회:학술대회논문집
• /
• 한국전기전자재료학회 2004년도 추계학술대회 논문집 Vol.17
• /
• pp.662-665
• /
• 2004

Recently, diagnosis techniques have been investigated to detect a partial discharge associated with a dielectric material defect in a high-voltage electrical apparatus However, the properties of detection technique of PD aren't completely understood because the physical process of PD. Therefore, this paper analyzes the process on Surface Discharge of Polymer Insulator using Wavelet transform. Wavelet transform provides a direct quantitative measure of spectral content in the time-frequency domain. As it is important to develop a non-contact method for detecting the Contamination Degree, this paper analyzes the electromagnetic waves emitted from PD using Wavelet transform. This paper experimentally shows the process of PD as a two-dimensional distribution in the time-frequency domain. This method is shown to be useful for detecting prediction of contamination degree.

• 한국지능시스템학회논문지
• /
• 제13권1호
• /
• pp.45-50
• /
• 2003

최근 인간형 로봇에 대한 개발이 괄목할 만한 성장을 이루고 있고, 친근한 로봇의 개발에 중요한 역할을 담당하는 것으로써 감성/감정의 인식이 필수적이라는 인식이 확산되고 있나. 본 논문은 음성의 감정인식에 있어 가장 큰 부분을 차지하는 피치의 패턴을 인식하여 감정을 분류/인식하는 시뮬레이터의 개발과 시뮬레이션 결과를 나타낸다. 또한, 피치뿐 아니라 음향학적으로 날카로움, 낮음 등의 요소를 분류의 기준으로 포함시켜서 좀더 신뢰성 있는 인식을 할 수 있음을 보인다. 주파수와 음성의 다양한 분석을 통하여, 음향적 요소와 감성의 상관관계에 대한 분석이 선행되어야 하므로, 본 논문은 사람들의 음성을 녹취하여 분석하였다 시뮬레이터의 내부 구조로는 음성으로부터 피치를 추출하는 부분과 피치의 패턴을 학습시키는 DRNN 부분으로 이루어져 있다.

• 음성과학
• /
• 제15권2호
• /
• pp.67-77
• /
• 2008

In this paper, we implemented a robust speech recognizer using the TMS320VC33 DSP. For this implementation, we had built speech and noise database suitable for the recognizer using spectral subtraction method for noise removal. The recognizer has an explicit structure in aspect that a speech signal is enhanced through spectral subtraction before endpoints detection and feature extraction. This helps make the operation of the recognizer clear and build HMM models which give minimum model-mismatch. Since the recognizer was developed for the purpose of controlling car facilities and voice dialing, it has two recognition engines, speaker independent one for controlling car facilities and speaker dependent one for voice dialing. We adopted a conventional DTW algorithm for the latter and a continuous HMM for the former. Though various off-line recognition test, we made a selection of optimal conditions of several recognition parameters for a resource-limited embedded recognizer, which led to HMM models of the three mixtures per state. The car noise added speech database is enhanced using spectral subtraction before HMM parameter estimation for reducing model-mismatch caused by nonlinear distortion from spectral subtraction. The hardware module developed includes a microcontroller for host interface which processes the protocol between the DSP and a host.

• 제어로봇시스템학회논문지
• /
• 제19권11호
• /
• pp.1022-1028
• /
• 2013

It is well knowns that based on the CSP (Common Spatial Pattern) algorithm, the linear projection of an EEG (Electroencephalography) signal can be made to spaces that optimize the discriminant between two patterns. Sharing disadvantages from linear time invariant systems, CSP suffers from the non-stationary nature of EEGs causing the performance of the classification in a BCI (Brain-Computer Interface) system to drop significantly when comparing the training data and test data. The author has suggested a simple idea based on the parallel model of CSP filters to improve the performance of BCI systems. The model was tested with a simple CSP algorithm (without any elaborate regularizing methods) and a perceptron learning algorithm as a classifier to determine the improvement of the system. The simulation showed that the parallel model could improve classification performance by over 10% compared to conventional CSP methods.