• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.13 no.1
• /
• pp.48-55
• /
• 2003

Hidden Markov Model(HMM) has a doubly embedded stochastic process with an underlying stochastic process that can be observed through another set of stochastic processes. This structure of HMM is useful for modeling vector sequence that doesn't look like a stochastic process but has a hidden stochastic process. So, HMM approach has become popular in various areas in last decade. The increasing popularity of HMM is based on two facts : rich mathematical structure and proven accuracy on critical application. In this paper, we applied continuous HMM (CHMM) approach with AR coefficient to detect and predict the chatter of lathe bite and to diagnose the wear of oil Journal bearing using rotor shaft displacement. Our examples show that CHMM approach is very efficient method for machine health monitoring and prediction.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.27 no.9
• /
• pp.1468-1475
• /
• 1990

In this paper, isolated words are recognized using multisection VQ and HMM. As recognition vocabuaries, 20 area-name which is uttered 5 times by 3 speakers is selected. In generating codebook, we devide recognition vocabulary into equal length, section, and make standard VQ codebook to each section and calculate observation by section and than recognize isolated words by HMM training. Multisection VQ codebook has time information and as observation is calculated by eacy section, computation is lesser and recongnition rate is higher than by whole codword. As a result, it is proved that recognition rate is higher in case of HMM using multisection VQ codebook.

• Phonetics and Speech Sciences
• /
• v.6 no.2
• /
• pp.29-34
• /
• 2014

In the Vector Taylor Series (VTS)-based noisy speech recognition methods, Hidden Markov Models (HMM) are usually trained with clean speech. However, better performance is expected by training the HMM with noisy speech. In a previous study, we could find that Minimum Mean Square Error (MMSE) estimation of the training noisy speech in the log-spectrum domain produce improved recognition results, but since the proposed algorithm was done in the log-spectrum domain, it could not be used for the HMM adaptation. In this paper, we modify the previous algorithm to derive a novel mathematical relation between test and training noisy speech in the cepstrum domain and the mean and covariance of the Multi-condition TRaining (MTR) trained noisy speech HMM are adapted. In the noisy speech recognition experiments on the Aurora 2 database, the proposed method produced 10.6% of relative improvement in Word Error Rates (WERs) over the MTR method while the previous MMSE estimation of the training noisy speech produced 4.3% of relative improvement, which shows the superiority of the proposed method.

• Journal of Communications and Networks
• /
• v.13 no.2
• /
• pp.125-131
• /
• 2011

Rapid strides are being made to bring to reality the technology of wearable sensors for monitoring patients' physiological data.We study the problem of automatically detecting anomalies in themeasured blood glucose levels. The normal daily measurements of the patient are used to train a hidden Markov model (HMM). The structure of the HMM-its states and output symbols-are selected to accurately model the typical transitions in blood glucose levels throughout a 24-hour period. The learning of the HMM is done using historic data of normal measurements. The HMM can then be used to detect anomalies in blood glucose levels being measured, if the inferred likelihood of the observed data is low in the world described by the HMM. Our simulation results show that our technique is accurate in detecting anomalies in glucose levels and is robust (i.e., no false positives) in the presence of reasonable changes in the patient's daily routine.

• KSCI Review
• /
• v.15 no.1
• /
• pp.117-124
• /
• 2007

통계 기반 접근 방법을 이용한 품사태깅에서 태깅 정확도는 훈련 데이터의 양에 좌우될 뿐 아니라, 말뭉치가 충분할지라도 수작업으로 구축한 말뭉치의 경우 항상 오류의 가능성을 내포하고 있으며 언어의 특성상 통계적으로 신뢰할만한 데이터의 수집에도 어려움이 따른다. 훈련 데이터로 사용되는 말뭉치는 많은 사람들이 수작업으로 구축하므로 작업자 중 일부가 언어에 대한 지식이 부족하다거나 주관적인 판단에 의한 태깅 실수를 포함할 수도 있기 때문에 단순한 저빈도와 관련된 잡음 외의 오류들이 포함될 수 있는데 이러한 오류들은 재추정이나 평탄화 기법으로 해결될 수 있는 문제가 아니다. 본 논문에서는 HMM(Hidden Markov Model)을 이용한 한국어 품사 태깅에서 재추정 후 여전히 존재하는 말뭉치의 잡음에 인한 태깅 오류 해결을 위해 비터비 알고리즘적용 단계에서 데이터 부족과 말뭉치의 오류로 인해 문제가 되는 부분을 찾아내고 규칙을 통해 수정을 하여 태깅 결과를 개선하는 방안을 제안한다. 실험결과는 오류가 존재하는 말뭉치를 사용하여 구현된 HMM과 비터비 알고리즘을 적용한 태깅 정확도에 비해 오류를 수정하는 과정을 거친 후 정확도가 향상됨을 보여준다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2004.05b
• /
• pp.587-590
• /
• 2004

In this paper proposed a HMM(Hidden Martov Model)-based segmentation method which is able to model shadows as well as foreground and background regions. Shadow of moving objects often obstruct visual tracking. We propose an HMM-based segmentation method which classifies in real time oath objects. In the case of traffic monitoring movies, the effectiveness of the proposed method has been proven through experimental results

• International journal of advanced smart convergence
• /
• v.13 no.1
• /
• pp.69-76
• /
• 2024

The objective of this paper is to investigate digital transformation in logistics within the shipping and logistics industry. This paper examines the case of Hyundai Merchant Marine (HMM) and draws insights from its success in becoming the top company amidst the industry challenges. To achieve a more in-depth analysis, a single-case study approach was employed. Various secondary data have been analyzed to support the examination of HMM's digital transformation. HMM has emerged as a global leader in transportation services. The company's success is attributed to its development of diverse ships, robust logistics systems, and pioneering a digital logistics platform, ultimately overcoming the crisis. Especially, this paper show that the digitalization of logistics has been analyzed as the most significant transformation. The findings of this paper can offer valuable implications for other companies grappling with similar challenges in the current trade landscape.

• Journal of Digital Convergence
• /
• v.13 no.8
• /
• pp.295-300
• /
• 2015

In vocabulary recognition using an HMM model which models the prior distribution for the observation of a discrete probability distribution indicates the advantages of low computational complexity, but relatively low recognition rate. The Bayesian techniques to improve vocabulary recognition model, it is proposed using a convergence of two methods to improve recognition noise-canceling recognition. In this paper, using a convergence of the prior probability method and techniques of Bayesian posterior probability based on HMM remove noise and improves the recognition rate. The result of applying the proposed method, the recognition rate of 97.9% in vocabulary recognition, respectively.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.12 no.2
• /
• pp.119-126
• /
• 2012

Recently, we used to recognize for human motions with some recognition algorithms. examples, HMM, DTW, PCA etc. In many human motions, we concentrated our research on recognizing fighting motions. In previous work, to obtain the fighting motion data, we used motion capture system which is developed with some active markers and infrared rays cameras and 3 dimension information converting algorithms by the stereo matching method. In this paper, we describe that the different method to acquiring 3 dimension fighting motion data and a HMM algorithm to recognize the data. One of the obtaining 3d data we used is depth map algorithm which is calculated by a stereo method. We test the 3d acquiring and the motion recognition system, and show the results of accuracy and performance results.

• The Journal of the Acoustical Society of Korea
• /
• v.19 no.6
• /
• pp.41-46
• /
• 2000

In this paper, we implement the speech recognition system based on the continuous distribution HMM and Browser-embedded model using the Java. That is developed for the speech analysis, processing and recognition on the Web. Client sends server through the socket to the speech informations that extracting of end-point detection, MFCC, energy and delta coefficients using the Java Applet. The sewer consists of the HMM recognizer and trained DB which recognizes the speech and display the recognized text back to the client. Because of speech recognition system using the java is high error rate, the platform is independent of system on the network. But the meaning of implemented system is merged into multi-media parts and shows new information and communication service possibility in the future.