• The Journal of the Korea institute of electronic communication sciences
• /
• v.9 no.11
• /
• pp.1221-1226
• /
• 2014

In an effort to improve the efficiency of the speech recognition, we investigate the effect of the number of phoneme clusters. For this purpose, codebooks of varied number of phoneme clusters are prepared by modified k-means clustering algorithm. The subsequent processing is fuzzy vector quantization (FVQ) and hidden Markov model (HMM) for speech recognition test. The result shows that there are two distinct regimes. For large number of phoneme clusters, the recognition performance is roughly independent of it. For small number of phoneme clusters, however, the recognition error rate increases nonlinearly as it is decreased. From numerical calculation, it is found that this nonlinear regime might be modeled by a power law function. The result also shows that about 166 phoneme clusters would be the optimal number for recognition of 300 isolated words. This amounts to roughly 3 variations per phoneme.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2003.06a
• /
• pp.6-9
• /
• 2003

In this paper, we studied on the difference of speaker adaptation according to the phoneme classification for Korean Speech recognition. In order to study of speech adaptation according to the weight of difference of phoneme as recognition unit, we used SCHMM as recognition system. And Speaker adaptation method used in this paper was MAPE(Maximum A Posteriori Probability Estimation), Linear Spectral Estimation. In order to evaluate the performance of these methods, we used 10 Korean isolated numbers as the experimental data. It is possible for the first and the second methods to be carried out unsupervised learning and used in on-line system. And the first method was shown performance improvement over the second method, and hybrid adaptation showed the better recognition results than those which performed each method. And the result of Speaker adaptation using the variable weight according to the phoneme had better than the result using fixed weight.

• Journal of KIISE:Information Networking
• /
• v.30 no.3
• /
• pp.397-406
• /
• 2003

As the information technology grows interests in the intrusion detection system (IDS), which detects unauthorized usage, misuse by a local user and modification of important data, has been raised. In the field of anomaly-based IDS several artificial intelligence techniques such as hidden Markov model (HMM), artificial neural network, statistical techniques and expert systems are used to model network rackets, system call audit data, etc. However, there are undetectable intrusion types for each measure and modeling method because each intrusion type makes anomalies at individual measure. To overcome this drawback of single-measure anomaly detector, this paper proposes a multiple-measure intrusion detection method. We measure normal behavior by systems calls, resource usage and file access events and build up profiles for normal behavior with hidden Markov model, statistical method and rule-base method, which are integrated with a rule-based approach. Experimental results with real data clearly demonstrate the effectiveness of the proposed method that has significantly low false-positive error rate against various types of intrusion.

• Korean Journal of Food Science and Technology
• /
• v.18 no.3
• /
• pp.226-233
• /
• 1986

Myofibrillar proteins were prepared from red muscle and white muscle, and their thermostabilities were compared. Rate constants of inactivation of myofibrillar proteins were increased as the ionic strength of reaction mixture increased and also dielectric constant of reaction mixture decreased. Thermodynamic data forinactivation of myofibrillar proteins, such as $D-value,\;{\Delta}H^{\ddag},\;{\Delta}G{\ddag}\;and\;{\Delta}S^{\ddag$, revealed that thermostabilities of myofibrillar proteins from white muscle were higher than those from red muscle, and that myofibrillar proteins from chicken muscle were more heatlabile than from bovine muscle.

• Korean Journal of Food Science and Technology
• /
• v.18 no.3
• /
• pp.181-186
• /
• 1986

Myosin were prepared from red muscle and white muscle, and their ATPase activities were compared. Ca-ATPase activity of bovine myosin from red muscle was higher than that of myosin from white muscle, while Ca-ATPase activity of chicken myosin from red muscle differed hardly from that of myosin from whitemuscle. Atso EDTA-ATPase activity of bovine red muscle myosin was higher than that of white muscle myosin ,although EDTA-ATPase activity of chicken myosin from red muscle differed hardly from that of white muscle myosin. When myosins were treated with trypsin, bovine myosin from white muscle was hydrolysed moreeasily than red muscle myosin was. Chicken myosin from red muscle , however, was hydrolysed by trypsin more easily than white muscle myosin was.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2017.05a
• /
• pp.83-83
• /
• 2017

수자원공학에서 일강수량 모의기법은 다양한 목적으로 활용되고 있지만, 일반적으로 홍수와 가뭄의 영향을 고려할 수 있는 수공구조물의 위험도 및 신뢰성 평가 및 수자원 계획을 수립하기 위한 입력 자료생성을 목적으로 활용된다. 유역 단위의 분석시 단일 지점에 대한 강수 모의 기법을 적용할 경우 각각의 지점에서 관측된 강수 자료의 시계열 및 통계치 특성이 효과적으로 재현되지만 공간적으로 발생하는 즉, 지점 간의 종속관계를 재현하지 못하는 문제가 발생한다. 이러한 이유로 공간적인 전이 특성이 있는 가뭄 분석 및 유역내 유출량의 공간적 변동 특성 분석에 단일지점별 모의 결과를 이용할 경우 관측 자료와 상반된 공간적 변동성으로 인하여 잘못된 가뭄 및 유출 분석 결과가 도출되는 문제점이 있다. 따라서, 실제적으로 발생하는 강수 특성을 반영한 유역 단위의 홍수 및 가뭄 등의 수문 분석을 위해서는 지점간의 종속성을 반영할 수 있는 다지점 강수 모의 모형의 적용이 필수적이다. 본 연구에서는 다지점 모의에 있어서, Wilks 모형의 지점별 시변동 특성과 공간상관성 재현 능력, HMM 모형이 갖는 강수 사상별로 분포된 양적 분포 패턴 재현 능력을 복합적으로 나타낼 수 있는 새로운 다지점 일강수량 모의 모형인 기계학습 기반 범주화 기법을 이용한 다지점 일강수량 모의 모형(ML-MRS)을 개발하였다. 또한, 지점별 강수량에 적용되는 확률분포모형은 Gamma 분포로 구성된 혼합모형을 적용하여 단일 확률 분포 모형의 자료 적합 문제를 개선하였다. 모의를 통한 일강수량 시계열 자료는 일 강수자료의 통계량을 효과적으로 모의하였으며, 다지점 모형의 모의 결과를 적용한 가뭄 모의 결과 관측 자료에서 나타나는 공간적 패턴이 재현되었다. 본 모형은 시 공간적 사상을 효과적으로 재현함으로서 지역의 변동특성을 반영한 가뭄, 홍수, 기상 현상 분석 등 활용도가 매우 높을 것으로 판단된다.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.16 no.1
• /
• pp.105-114
• /
• 2006

Ubiquitous Sensor Network consists of a number of sensor nodes with a limited computation power and limited communication capabilities, and a sensor node is able to communicate with each other at anytime and in any place. Due to the rapid research and development in sensor networks, it will rapidly grow into environments where hmm beings can interact in an intuitive way with sensing objects which can be PDAs, sensors, or even clothes in the future. We are aiming at realizing an Unattended Secure Security System to apply it to Ubiquitous Sensor Network. In this paper, the vulnerabilities in the Unattended security system are identified, and a new protocol called OMAC-SNEP is proposed for the Unattended Secure Security System. Because the CBC-MAC in SNEP is not secure unless the message length is fixed, the CBC-MAC in SNEP was replaced with OMAC in SNEP. We have shown that the proposed protocol is secure for my bit length of messages and is almost as efficient as the CBC-MAC with only one key. OMAC-SNEP can be used not only in Unattended Security System, but also any other Sensor Networks.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.16 no.6
• /
• pp.81-93
• /
• 2006

A file system is an evidence resource of cyber crime in computer forensics. Therefore the methods of recovering the file system and searching important information have been offered. However, the methods for finding a malicious fie in free blocks or slack spaces have not been suggested. In this paper, we propose an investigation method to find a maliciously executable fragmented file. After estimating if a file is executable with a machine code rate, we conclude it could be malicious by comparing a similarity of instruction sequences. To examine instruction sequences, we also propose a method of profiling malicious files using file and a method of comparing the continued scores. As the results, we could exactly pick out the malicious execution files, such as buffer overflow attack program, at fitting threshold level.

• Journal of the Society of Naval Architects of Korea
• /
• v.58 no.6
• /
• pp.366-374
• /
• 2021

This paper presents the analysis of the speed-power performance in the real sea using a large container vessel data provided as a test bed from a shipping company. To perform a speed trial of the vessel during a commercial voyage, the on-board measuring device and various operation data acquisition systems were mounted on the vessel for long-term performance monitoring and the voyage operated under the container loading condition close to the design draft was adopted. The content of this paper consists of Part I and Part II. Part I, such as this paper, contains the speed trial method and analysis results of the operating vessel. Part II contains the analysis of the speed-power characteristics change over time and before and after hull cleaning using operation data measured from the voyage operated under a condition similar to the speed trial.

• Nuclear Engineering and Technology
• /
• v.54 no.11
• /
• pp.4111-4124
• /
• 2022

The control rod drive system is critical to the reactor's reliable operation. The performance of its control system and mechanical system will gradually deteriorate because of operational and environmental stresses, thus increasing the reactor's operational risk. Currently there are few researches on the aging-related degradation of the entire control rod drive system. Because it is difficult to quantify the effect of various environmental stresses and establish an accurate physical model when multiple mechanisms superimposed in the degradation process. Therefore, this paper investigates the aging-related degradation of a control rod drive system by integrating Dynamic Object-Oriented Bayesian Network and Hidden Markov Model. Uncertainties in the degradation of the control system and mechanical system are addressed by using fuzzy theory and the Hidden Markov Model respectively. A system which consists of eight control rod drive mechanisms divided into two groups is used to demonstrate the method. The aging-related degradation of the control rod drive system is analyzed by the Bayesian inference algorithm based on the accelerated life test data, and the impact of different operating schemes on the system performance is also investigated. Meanwhile, the components or units that have major impact on the system's performance are identified at different operational phases. Finally, several essential safety measures are suggested to mitigate the risk caused by the system degradation.