• 정보보호학회논문지
• /
• 제29권6호
• /
• pp.1393-1401
• /
• 2019

최근 시스템에 음성 인증 기능이 탑재됨에 따라 화자(Speaker)를 정확하게 인증하는 중요성이 높아지고 있다. 이에 따라 다양한 방법으로 화자를 인증하는 모델이 제시되어 왔다. 본 논문에서는 Short-time Fourier transform(STFT)를 적용한 새로운 화자 인증 모델을 제안한다. 이 모델은 기존의 Mel-Frequency Cepstrum Coefficients(MFCC) 추출 방법과 달리 윈도우 함수를 약 66.1% 오버랩하여 화자 인증 시 정확도를 높일 수 있다. 새로운 화자 인증 모델을 제안한다. 이 때, LSTM 셀을 적용한 Recurrent Neural Network(RNN)라는 딥러닝 모델을 사용하여 시변적 특징을 가지는 화자의 음성 특징을 학습하고, 정확도가 92.8%로 기존의 화자 인증 모델보다 5.5% 정확도가 높게 측정되었다.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2005년도 춘계종합학술대회
• /
• pp.493-496
• /
• 2005

본 논문에서는 번호판 고유의 복합 색상 정보와 수직 에지 정보를 이용한 번호판 후보 영역 추출기법을 제안한다. 또한 추출된 번호판 영역의 정확성을 높이기 위해서, Fast DCT를 거쳐 압축된 이미지에 대하여 Support Vector Machines(SVM)을 이용한 검증 과정을 제안한다. 제안하는 기법은 차량번호판 인식의 대상이 되는 자동차의 위치가 정면, 후면을 구분하지 않는 다양함을 가지고, 주변 배경이 충분히 포함되는 상황에서 다양한 크기를 가지는 355장의 영상들을 대상으로 한다. 실험 결과, SVM을 이용한 검증 과정을 거친 방법이 그렇지 않은 경우보다 20%이상 향상된 번호판 추출 성공률을 나타내었다.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2004년도 춘계학술대회 논문집
• /
• pp.247-252
• /
• 2004

Developed shape awareness technology and vision technology for optical ten slant in this research and including external form state of lens for the performance verification developed so that can be good achieve badness finding. And, establish to existing reflex data because inputting surface badness degree of scratch's standard specification condition directly, and error designed to distinguish from product more than schedule error to badness product by normalcy product within schedule extent after calculate the error comparing actuality measurement reflex data md standard reflex data mutually. Developed system to smallest 1pixel unit though measuring is possible 1pixel as 3.7$\mu\textrm{m}$${\times}$3.7$\mu\textrm{m}$(0.1369${\times}$10/sub-1/$\textrm{mm}^2$) the accuracy to 10/sub-1/mm minutely measuring is possible performance verification and trust ability through an experiment prove.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권1호
• /
• pp.168-182
• /
• 2020

Spoofing attacks, especially replay attacks, pose great security challenges to automatic speaker verification (ASV) systems. Current works on replay attacks detection primarily focused on either developing new features or improving classifier performance, ignoring the effects of feature variability, e.g., the channel variability. In this paper, we first establish a mathematical model for replay speech and introduce a method for eliminating the negative interference of the channel. Then a novel feature is proposed to detect the replay attacks. To further boost the detection performance, four post-processing methods using normalization techniques are investigated. We evaluate our proposed method on the ASVspoof 2017 dataset. The experimental results show that our approach outperforms the competing methods in terms of detection accuracy. More interestingly, we find that the proposed normalization strategy could also improve the performance of the existing algorithms.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2005년도 춘계학술대회 논문집
• /
• pp.1591-1596
• /
• 2005

In this paper, we designed the aspheric lens by mathematical methods which are solved by refractive index. Based on these mathematical method, we manufactured the aspheric lens. Because of the usefulness of manufacturing, we used the acryl as the material. And we used a high speed machine to manufacture the aspheric lens. Also it sits in judgment the aberration of spherical distribution by using the grid fringe. Asource of laser produced by laser sheet generator passed the optical system of the aspheric lens. Though a character is to verify that the accuracy of the aspheric lens by experiment of the straight character.

• 한국원자력학회:학술대회논문집
• /
• 한국원자력학회 1996년도 춘계학술발표회논문집(2)
• /
• pp.499-504
• /
• 1996

Current Probabilistic Safety Assessment (PSA) techniques are not usually utilized for day-to-day applications in nuclear power plants. The major reason for this anomaly is the complexity of plant models developed for PSA studies and the multitude of resulting fault trees. This impediment can be overcome by the use of simplified plant models. However, oversimplified models usually result in loss of valuable information and therefore. simplification approaches have to be used judiciously in order to achieve accurate and meaningful results. For this reason. development of an appropriate simplification approach must be performed using extreme caution followed with results verification in sequence as well as system levels. If there are no significant differences between the simplified and the original models, the simplified model can be efficiently used in the application of the PSA. This paper presents a methodology for how to develop a suitable simplification technique and the results of its verification for sample systems and sequences. The results show that the utilization of simplified plant models will significantly reduce the number of fault trees with no significant loss of accuracy.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2006년도 추계학술대회논문집
• /
• pp.519-522
• /
• 2006

Accurate modeling of a dynamic system from experimental data is the bases for the model updating or heath monitoring of the system. Modal analysis or modal test is a routine process to get the modal parameters of a dynamic system. The modal parameters include the natural frequencies, damping ratios and mode shapes. This paper presents a new method that can derive the equations of motion for a dynamic system from the modal parameters obtained by the modal analysis or modal test. The present method based on the relation between the eigenvalues and eigenvectors of the state space equation derives the mass, damping and stiffness matrices of the system. The modeling of a cantilevered beam from modal parameters is an example to prove the efficiency and accuracy of the present method. Using the lateral displacements only, not the rotations, gives limited information for the system. The numerical verification up to now gives reasonable results and the verification with the test data is scheduled.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2005년도 제36회 하계학술대회 논문집 A
• /
• pp.795-797
• /
• 2005

This paper describes a measurement & verification scheme for the high efferent lighting program in Demand Side Management (DSM), which was Introduced in the beginning of the 1980's in Korea With the dramatic changes in the circumstance of power industry in Korea. It is required the reviewing for the existing process of DSM programs and the development of the new process methods/standards on the evaluation of DSM programs. To design an accurate evaluation scheme for the high efficient lighting program. In this paper, we develop an evaluation prototype for the high efficient lighting program using measurements and stipulated factors. The proposed scheme is tested by practical measurements and modified stipulated factors Results from the case studios show the proposed scheme could provide the practical performance evaluation results with promising accuracy.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1997년도 추계학술대회 논문집
• /
• pp.425-428
• /
• 1997

Effectiveness of corrective machining algorithm is verified experimentally in this paper by performing corrective lapping work to single side and double sides hydrostatic tables. Lapping is applied as machining method. Machining information is calculated from measured motion errors by applying the algorithm, without information on rail profile. It is possible to acquire 0.13pm of linear motion error, 1.40arcsec of angular motion error in the case of single side table, and 0.07pm of linear motion error, 1.42arcsec of angular motion error in the case of double sides table. The experiment is performed by the unskilled person after he experienced a little of preliminary machining. Experimental results show that corrective machining algorithm is very effective, and anyone can improve the accuracy of hydrostatic table by using the algorithm.

• 한국정밀공학회지
• /
• 제8권4호
• /
• pp.84-94
• /
• 1991

In order to achieve high accuracy of teaching and increase productivity using industrial robots in polishing process of dies, an off-line task programming system was developed on IBM-PC/386 under WINDOWS 3.0 operating system. The internal structure and the machematical basis of CAMPoli are described. Surface modeling technique of polishing dies with sculptured surfaces is introduced by poing data interpolation methodology through the use of CL-data transmitted from conventional CAM system. Tool selection, polishing speed, polishing pressure and kinds of tool motions can be determined and selected by user specified polishing variables. Task creation and verification of polishing path via computer graphics simulation of polishing tool can be done by the menu- driven function of CAMPoli system. Post-processing module is attached to generate robot language. Some simulation results are provided as verification means of the system.