• Journal of Sensor Science and Technology
• /
• v.16 no.5
• /
• pp.369-376
• /
• 2007

Recently, obesity that is increasingly becoming a major cause of various diseases is emerging as a serious social problem. In order to solve this problem, the necessity of measurement systems for overweight management has increased. This paper is a study on the measurement system for obesity management that can offer right medical services everywhere and allways by analyzing EMG (electromyograph) of the abdomen and then checking one's health state. For analyzing EMG signals of the abdomen, algorithms for energy detection, signal feature extraction, classification and recognition are presented. This paper proposes a system that provides an appropriate an estimation on the health status by evaluating the obesity degree and muscular strength of the abdomen through the system applying these algorithms.

• Journal of the Korean Society of Laryngology, Phoniatrics and Logopedics
• /
• v.8 no.2
• /
• pp.173-177
• /
• 1997

The speech movements are acquired activity, not determined by instincts or by biologic inheritance either. The child listens to the sound from the surrounding persons, observes the speech movement of the people and tried to imitate them. Then the child acquires their specific phonation pattern. We guessed that the parents influences to the child are very important in the developing of the speech movements. Because the parents are first contact person to the baby. The recognition of parents about the voice changes in the child will be important too. And social environments such as kindergarden, school, friends contact with, can influence to the voice of the child. We investigated the state of the voice, parents influence and social environmental factor. In the bases of this study, we knew that the parents recognition about the voice changes of child, faulty vocal habits of child, social environmental factors influenced to the voice of child. And we thought we have to do our best for the early detection of voice changes and proper treatment.

• Journal of Biosystems Engineering
• /
• v.24 no.1
• /
• pp.75-80
• /
• 1999

Signal separation and transmission are essential for automatic breeding management of domestic animals. Electronic identification system could transmit the signal of an individual within a defined range to a personal computer by an electromagnetic signal recognition method. Signals for individual recognition were originated by controlling 12 tri-state pins of IC(PT2262) in a transmitter. PT 2262 can generate 4,096 codes. These encoded signals were modulated and transmitted with wireless lines from the transmitter. Then they were demodulated in a receiver, and the signals were transmitted to the micro-processor through an interface and were identified in a PC.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2003.05b
• /
• pp.90-95
• /
• 2003

For the safety driving of an automobile which is become individual requisites, a new Neural Network algorithm which recognized the load vehicles in real time is proposed. The proposed neural network technique is the real time computation method through the inter-node diffusion. In the network, a node corresponds to a state in the quantized input space. Each node is composed of a processing unit and fixed weights from its neighbor nodes as well as its input terminal. The most reliable algorithm derived for real time recognition of vehicles, is a dynamic programming based algorithm based on sequence matching techniques that would process the data as it arrives and could therefore provide continuously updated neighbor information estimates. Through several simulation experiments, real time reconstruction of the nonlinear image information is processed. 1-D LIPN hardware has been composed and various experiments with static and dynamic signals have been implemented.

• International Journal of Railway
• /
• v.7 no.4
• /
• pp.94-99
• /
• 2014

Urban railway systems are located under populated areas and are mostly constructed for underground structures which demand high standards of structural safety. However, the damage progression of underground structures is hard to evaluate and damaged underground structures may not effectively stand against successive earthquakes. This study attempts to examine initial damage-stage and to access structural damage condition of the ground structures using Earthquake Damage Monitoring (EDM) system. For actual underground structure, vulnerable damaged member of Ulchiro-3ga station is chosen by finite element analysis using applied artificial earthquake load, and then damage pattern and history of damaged members is obtained from measured acceleration data introduced unsupervised learning recognition. The result showed damage index obtained by damage scenario establishment using acceleration response of selected vulnerable members is useful. Initial damage state is detected for selected vulnerable member according to established damage scenario. Stiffness degrading ratio is increasing whereas the value of reliability interval is decreasing.

• The Journal of the Acoustical Society of Korea
• /
• v.29 no.2E
• /
• pp.100-106
• /
• 2010

Multiple-model speech recognizer has been shown to be quite successful in noisy speech recognition. However, its performance has usually been tested using the general speech front-ends which do not incorporate any noise adaptive algorithms. For the accurate evaluation of the effectiveness of the multiple-model frame in noisy speech recognition, we used the state-of-the-art front-ends and compared its performance with the well-known multi-style training method. In addition, we improved the multiple-model speech recognizer by employing N-best reference HMMs for interpolation and using multiple SNR levels for training each of the reference HMM.

• Proceedings of the KIEE Conference
• /
• 1999.07g
• /
• pp.2821-2823
• /
• 1999

In this paper we proposed an algorithm for rotation-invariant pattern recognition and rotated angle estimation between two patterns by employing selective template matching. Generally template matching has been used in determining the location of pattern but template matching requires a number of calculating correlation. To reduce the number of correlation we used steady-state genetic algorithm which is effective in optimization problem. We apply this method to distinguish specific pattern from similar coin patterns and estimate rotated angle between patterns. Our result leads us to the conclusion that proposed method performed faster than classical template matching

• Journal of Information Processing Systems
• /
• v.13 no.4
• /
• pp.928-940
• /
• 2017

State-of-the-art speaker recognition systems may work better for the English language. However, if the same system is used for recognizing those who speak different languages, the systems may yield a poor performance. In this work, the decisions of a Gaussian mixture model-universal background model (GMM-UBM) and a learning vector quantization (LVQ) are combined to improve the recognition performance of a multilingual speaker identification system. The difference between these classifiers is in their modeling techniques. The former one is based on probabilistic approach and the latter one is based on the fine-tuning of neurons. Since the approaches are different, each modeling technique identifies different sets of speakers for the same database set. Therefore, the decisions of the classifiers may be used to improve the performance. In this study, multitaper mel-frequency cepstral coefficients (MFCCs) are used as the features and the monolingual and cross-lingual speaker identification studies are conducted using NIST-2003 and our own database. The experimental results show that the combined system improves the performance by nearly 10% compared with that of the individual classifier.

• Annual Conference of KIPS
• /
• 2018.10a
• /
• pp.613-616
• /
• 2018

Sequence tagging is an important task in Natural Language Processing (NLP), in which the Named Entity Recognition (NER) is the key issue. So far the most widely adopted model for NER in NLP is that of combining the neural network of bidirectional long short-term memory (BiLSTM) and the statistical sequence prediction method of Conditional Random Field (CRF). In this work, we improve the prediction accuracy of the BiLSTM by supporting an aligned word representation mechanism. We have performed experiments on multilingual (English, Spanish and Dutch) datasets and confirmed that our proposed model outperformed the existing state-of-the-art models.

• Proceedings of the KSPS conference
• /
• 2003.05a
• /
• pp.219-222
• /
• 2003

In this paper, we study the usage of phoneme duration information for rejection garbage sentence. First, we build a phoneme duration modeling in a speech recognition system based on dicicion tree state tying, We assume that phone duration has a Gamma distribution. Next, we build a verification module in which word-level confidence measure is used. Finally, we make a comparative study on phoneme duration with speech DB obtained from the live system. This DB consistes of OOT(out-of-task) and ING(in-grammar) utterences. the usage of phone duration information yields that OOT recognition rate is improved by 46％ and that another 8.4％ error rate is reduced when combined with utterence verification module.