• MALSORI
• /
• no.43
• /
• pp.179-188
• /
• 2002

Improving the recognition performance of connected digit telephone speech still remains a problem to be solved. As a basic study for it, this paper analyzes the variation of feature parameters of Korean digit telephone speech according to channel distortion. As a feature parameter for analysis and recognition MFCC is used. To analyze the effect of telephone channel distortion depending on each call, MFCCs are first obtained from the connected digit telephone speech for each phoneme included in the Korean digit. Then CMN, RTCN, and RASTA are applied to the MFCC as channel compensation techniques. Using the feature parameters of MFCC, MFCC+CMN, MFCC+RTCN, and MFCC+RASTA, variances of phonemes are analyzed and recognition experiments are done for each case. Experimental results are discussed with our findings and discussions

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.55 no.2
• /
• pp.75-81
• /
• 2006

The purpose of this paper is to recognize the feature points of ECG and human pulse -which signal shows the electric and physical characteristics of heart respectively- using wavelet transform. Wavelet transform is proper method to analyze a signal in time-frequency domain. In the process of wavelet decomposition and reconstruction of ECG and human pulse signal, we removed the noises of signal and recognized the feature points of signal using some of decomposed component of signal. We obtained the result of recognition rate that is estimated about 95.45$\%$ in case of QRS complex, 98.08$\%$ in case of S point and P point and 92.81$\%$ in case of C point. And we computed diagnosis parameters such as RRI, U-time and E-time.

• Journal of the Institute of Convergence Signal Processing
• /
• v.15 no.2
• /
• pp.37-41
• /
• 2014

The vector Taylor series (VTS) based method usually employs clean speech Hidden Markov Models (HMMs) when compensating speech feature vectors or adapting the parameters of trained HMMs. It is well-known that noisy speech HMMs trained by the Multi-condition TRaining (MTR) and the Multi-Model-based Speech Recognition framework (MMSR) method perform better than the clean speech HMM in noisy speech recognition. In this paper, we propose a method to use the noise-adapted HMMs in the VTS-based speech feature compensation method. We derived a novel mathematical relation between the train and the test noisy speech feature vector in the log-spectrum domain and the VTS is used to estimate the statistics of the test noisy speech. An iterative EM algorithm is used to estimate train noisy speech from the test noisy speech along with noise parameters. The proposed method was applied to the noise-adapted HMMs trained by the MTR and MMSR and could reduce the relative word error rate significantly in the noisy speech recognition experiments on the Aurora 2 database.

• Journal of the Korean Society for Precision Engineering
• /
• v.20 no.3
• /
• pp.105-111
• /
• 2003

This paper presents an approach to automate process planning of press dies for manufacturing of car bodies. Considering that the press-dies used at the same press operations regardless of the panels they produce or the car models of which they produce panels have similar shapes except for the forming part of the dies, general approaches to recognize manufacturing features from CAD models are not necessary. Therefore, a hybrid approach is proposed combining feature-based design and feature-extraction approaches. The proposed method recognizes features by parsing the parameters extracted from CAD models and finds proper operations by querying the database by the recognized features. An internet-based process planning system is developed to demonstrate the proposed approach and to suggest a new paradigm of process planning system that utilizes an internet access to the CAD system.

• Journal of Convergence for Information Technology
• /
• v.11 no.10
• /
• pp.10-19
• /
• 2021

In this paper, we tried to derive characteristic parameters that reflect mental fatigue through EEG measurement and analysis. For this purpose, mental fatigue was induced through a resting state with eyes closed and performing subtraction operations in mental arithmetic for 30 minutes. Five subjects participated in the experiment, and all subjects were right-handed male students in university, with an average age of 25.5 years. Spectral analysis was performed on the EEG collected at the beginning and the end of the experiment to derive feature parameters reflecting mental fatigue. As a result of the analysis, the absolute power of the alpha band in the occipital lobe and the temporal lobe increased as the mental fatigue increased, while the relative power decreased. Also, the difference in power between resting state and task state showed that the relative power was larger than the absolute power. These results indicate that alpha relative power in the occipital lobe and temporal lobe is a feature parameter reflecting mental fatigue. The results of this study can be utilized as feature parameters for the development of an automated system for mental fatigue determination such as fatigue and drowsiness while driving.

• The Journal of the Acoustical Society of Korea
• /
• v.31 no.1
• /
• pp.39-50
• /
• 2012

A voice transformation (VT) method that can make the utterance of a source speaker mimic that of a target speaker is described. Speaker individuality transformation is achieved by altering three feature parameters, which include the LPC cepstrum, pitch period and gain. The main objective of this study involves construction of an optimal sequence of features selected from a target speaker's database, to maximize both the correlation probabilities between the transformed and the source features and the likelihood of the transformed features with respect to the target model. A set of two-pass conversion rules is proposed, where the feature parameters are first selected from a database then the optimal sequence of the feature parameters is then constructed in the second pass. The conversion rules were developed using a statistical approach that employed a maximum likelihood criterion. In constructing an optimal sequence of the features, a hidden Markov model (HMM) was employed to find the most likely combination of the features with respect to the target speaker's model. The effectiveness of the proposed transformation method was evaluated using objective tests and informal listening tests. We confirmed that the proposed method leads to perceptually more preferred results, compared with the conventional methods.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.60 no.5
• /
• pp.1055-1061
• /
• 2011

In this study, main parameters: aperture, first bend and second bend which express a structure of ear canal are extracted in order to modeling and manufacture the ready-made ear shells of hearing aids. The proposed parameter extraction method consists of 2 important algorithms, aperture detection and feature detection. In the aperture detection algorithm, aperture of 3-D scanned virtual ear impression and parameters relating to ear shell of hearing aid are determined. The feature detection algorithm detects first bend, second bend, and related parameters. Through these two algorithms, parameters for aperture, first bend, and second bend are extracted to model the ready-made ear shell of hearing aid. The values of these extracted parameters from 36 people's right ear impression are analyzed and measured statistically. As a result of the analysis, it has been found that it is possible to classify ready-made ear shell parameters by age and size. The ready-made ear shell parameters are classified 3-size for 20 years old and 2-size for 60 years olde. Using 3D rhino program, virtual ready-made ear shell is reconstructed by parameters of every type, and simulated to model it. A final product was produced by transferring simulation result with rapid prototyping system. The modeled ready-made ear shell is evaluated with the objective and subjective method. Objective method is the comparison volume ratio and overlapped volume ratio of ear impression from randomly chosen 18 people and ready-made ear shell. And subjective method is that the final product of ready-made ear shell is used by users and the satisfaction number drawn from well fitting and comfortable testing was evaluated. In the result of the evaluation, it has been found that volume ration is 70%, big and middle size ready-made ear shell products are possible, and the satisfaction number is high.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.2E
• /
• pp.98-104
• /
• 2002

The present study describes a combination method to recognize the human affective states such as anger, happiness, sadness, or surprise. For this, we extracted emotional features from voice signals and facial expressions, and then trained them to recognize emotional states using hidden Markov model (HMM) and neural network (NN). For voices, we used prosodic parameters such as pitch signals, energy, and their derivatives, which were then trained by HMM for recognition. For facial expressions, on the other hands, we used feature parameters extracted from thermal and visible images, and these feature parameters were then trained by NN for recognition. The recognition rates for the combined parameters obtained from voice and facial expressions showed better performance than any of two isolated sets of parameters. The simulation results were also compared with human questionnaire results.

• Proceedings of the IEEK Conference
• /
• 1999.06a
• /
• pp.555-558
• /
• 1999

In this paper, Ire present the result of feature points recognition and classification of radial pulse by the shape of pulse wave. The recognition algorithm use the method which runs in parallel with both the data of ECG and differential pulse simultaneously to recognize the feature points. Also we specified 3-time elements of pulse wave as main parameters for diagnosis and measured them by execution of algorithm. then we classify the shape of radial pulse by existence and position of feature points.

• 제어로봇시스템학회:학술대회논문집
• /
• 1993.10b
• /
• pp.18-23
• /
• 1993

In this paper, we attempt to estimate the state of a finite state system. In such system, we can observe time series data which has some significant behaviors corresponding to its system states. The behavior is characterized by feature parameters extracted from time series. Our thought is that the system output time series data is expressed as a sequence of behavior patterns which are represented by clusters in feature parameters space. An algorithm jointing fuzzy clustering to fuzzy finite state transition model is suggested.