• The Journal of the Acoustical Society of Korea
• /
• v.15 no.2
• /
• pp.40-48
• /
• 1996

In this paper, a 40 dimensional segment vector with 4 frame and 7 frame width in every monosyllable interval was compressed into a 10, 14, 20 dimensional vector using K-L expansion and neural networks, and these was used to speech recognition feature parameter for CHMM. And we also compared them with CHMM added as feature parameter to the discrete duration time, the regression coefficients and the mixture distribution. In recognition test at 100 monosyllable, recognition rates of CHMM +${\bigtriangleup}$MCEP, CHMM +MIX and CHMM +DD respectively improve 1.4%, 2.36% and 2.78% over 85.19% of CHMM. And those using vector compressed by K-L expansion are less than MCEP + ${\bigtriangleup}$MCEP but those using K-L + MCEP, K-L + ${\bigtriangleup}$MCEP are almost same. Neural networks reflect more the speech dynamic variety than K-L expansion because they use the sigmoid function for the non-linear transform. Recognition rates using vector compressed by neural networks are higher than those using of K-L expansion and other methods.

• Journal of Biomedical Engineering Research
• /
• v.31 no.4
• /
• pp.321-327
• /
• 2010

Most of the ear shells of hearing aids are manufactured manually, and it is one of the reasons that the cost of the custom-made hearing aids can be increased. Thus it is required to manufacture the ready-made ear shell for the purpose of easy manufacturing and decrease in cost. In this study, we extract parameters in order to manufacture the ready-made ear shell for CIC type hearing aids and simulate to reconstruct the ear shell using the extracted parameters. To parameter extraction, we set up the eleven parameters for the ready-made ear shell based on anatomical characteristics of the ear canal, and we found values of the parameters from twenty-one impressions in their 20s and twelve impressions in their 60s using aperture detection and feature detection algorithms. Classifying the parameters by size, we also determine the parameters of ready-made ear shell into three types for people in their 20s and two types for people in their 60s. Each ready-made ear shell was simulated to reconstruct using figured parameters, and evaluated the rate of agreement with unused impressions for setting parameters. To evaluate the ready-made ear shell, we calculate the volume ratio and intersection between of the each impression and ready-made ear shell, and the intersection ratio using the intersection volume and ready-made ear shell volume. As a result, the volume ratio was about 70%, and volume match ratio was also up to 70%. It means that the ready-made ear shell we simulated is the significantly matched to impression.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.11
• /
• pp.2137-2142
• /
• 2016

The electrical current generated by heart creates not only electric potential but also a magnetic field. In this study, the signals obtained magnetocardiogram(MCG) using 61 channel superconducting quantum interference device(SQUID) system, and the clinical significance of various feature parameters has been developed MCG. Neural network algorithm was used to perform the classification of ischemic heart disease. The MCG signal was obtained to facilitate the extraction of parameters through a process of pre-processing. The data used to research the normal group 10 and ischemic heart disease group 10 with visible signs of stable angina patients. The available clinical indicators were extracted by characteristic point, characteristic interval parameter, and amplitude ratio parameter. The extracted parameters are determined to analysis the significance and clinical parameters were defined. It is possible to classify ischemic heart disease using the MCG feature parameters as a neural network input.

• The Journal of Society for e-Business Studies
• /
• v.23 no.3
• /
• pp.225-237
• /
• 2018

The demand for processing large data streams is growing rapidly as the generation and processing of large volumes of data become more popular. A variety of large data processing technologies are being developed to suit the increasing demand. One of the technologies that researchers have particularly observed is the data stream clustering with sliding windows. Data stream clustering with sliding windows may create a new set of clusters whenever the window moves. Previous data stream clustering techniques with sliding windows exploit the coresets, also known as group features that summarize the data. In this paper, we present some reformable elements of a group-feature based algorithm, and propose our algorithm that modified the clustering algorithm of the original one. We conduct a performance comparison between two algorithms by using different parameter values. Finally, we provide some guideline for the selective use of those algorithms with regard to the parameter values and their impacts on the performance.

• Smart Media Journal
• /
• v.12 no.5
• /
• pp.28-35
• /
• 2023

In this study, we propose a new light-weight model RoutingConvNet with fewer parameters to improve the applicability and practicality of speech emotion recognition. To reduce the number of learnable parameters, the proposed model connects bidirectional MFCCs on a channel-by-channel basis to learn long-term emotion dependence and extract contextual features. A light-weight deep CNN is constructed for low-level feature extraction, and self-attention is used to obtain information about channel and spatial signals in speech signals. In addition, we apply dynamic routing to improve the accuracy and construct a model that is robust to feature variations. The proposed model shows parameter reduction and accuracy improvement in the overall experiments of speech emotion datasets (EMO-DB, RAVDESS, and IEMOCAP), achieving 87.86%, 83.44%, and 66.06% accuracy respectively with about 156,000 parameters. In this study, we proposed a metric to calculate the trade-off between the number of parameters and accuracy for performance evaluation against light-weight.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.1
• /
• pp.36-42
• /
• 2006

A Principal component analysis (PCA)-based Mel-Frequency Discrete Wavelet Coefficients (MFDWC) feature Parameters for speaker verification system is Presented in this Paper In this method, we used the 1st-eigenvector obtained from PCA to calculate the energy of each node of level that was approximated by. met-scale. This eigenvector satisfies the constraint of general weighting function that the squared sum of each component of weighting function is unity and is considered to represent speaker's characteristic closely because the 1st-eigenvector of each speaker is fairly different from the others. For verification. we used Universal Background Model (UBM) approach that compares claimed speaker s model with UBM on frame-level. We performed experiments to test the effectiveness of PCA-based parameter and found that our Proposed Parameters could obtain improved average Performance of $0.80\%$compared to MFCC. $5.14\%$ to LPCC and 6.69 to existing MFDWC.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2008.05a
• /
• pp.505-508
• /
• 2008

Electrocardiogram (ECG) and blood pressure (BP) are main vital signs which are the standards in most medical settings in assessing the most basic body functions. Multi parameters are desired in providing more information for health professionals in order to detect or monitor medical problems of patients more precisely. This study urges us to develop a robust wireless healthcare monitoring system which has multiple physiological signs measurements on real time that applicable to various environments which integrates wireless sensor network technology and code division multiple access (CDMA) network with extended feature of locally standalone diagnosis algorithms that implemented in tell phone. ECG signal and BP parameter of the patients are routinely be monitored, processed and analyzed in details at cell phone locally to produce useful medical information to ease patients for tracking and future reference purposes. Any suspected or unknown patterns of signals will be immediately forwarded to hospital server using cell phone for doctors' evaluation. This feature enables the patients always recognize the importance of self-health checking so that the preventive actions can be taken earlier through this analytic information provided by this monitoring system because "Prevention is better than Cure".

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.41 no.1
• /
• pp.9-16
• /
• 2004

In this paper, we propose a robust controller for robot manipulators with parametric uncertainties using feature-based visual servo control system. In order to improve trajectory error of the robot manipulators due to the parameter variation, integral action is included in the dynamic control of part in inner subroutine of the control system. This integral action also reduces feature error in the steady state. The stability analysis of the closed-loop system is shown by the Lyapunov method. The effectiveness of the proposed method is shown by simulation and experimental results on the 5 link robot manipulator with two degree of freedom.

• Journal of the Korea Society of Computer and Information
• /
• v.22 no.9
• /
• pp.65-71
• /
• 2017

This paper proposes a smart phone authentication method based on user's behavior and habit that is an authentication method against shoulder surfing attack and brute force attack. As smart phones evolve not only storage of personal data but also a key means of financial services, the importance of personal information security in smart phones is growing. When user authentication of smart phone, pattern authentication method is simple to use and memorize, but it is prone to leak and vulnerable to attack. Using the features of the smart phone pattern method of the user, the pressure applied when touching the touch pad with the finger, the size of the area touching the finger, and the time of completing the pattern are used as feature vectors and applied to user authentication security. First, a smart phone user models and stores three parameter values as prototypes for each section of the pattern. Then, when a new authentication request is made, the feature vector of the input pattern is obtained and compared with the stored model to decide whether to approve the access to the smart phone. The experimental results confirm that the proposed technique shows a robust authentication security using subjective data of smart phone user based on habits and behaviors.

• Journal of Biomedical Engineering Research
• /
• v.25 no.3
• /
• pp.217-225
• /
• 2004

In this paper, a new automatic classification method for the normal EEC and schizophrenia EEC using hidden Markov model(HMM) is proposed. We used the feature parameters which are the variance for statistical stationary interval of the EEC and power spectrum ratio of the alpha, beta, and theta wave. The results were shown that high classification accuracy of 90.9％ in the case of normal person, and 90.5％ in the case of schizophrenia patient. It seems that proposed classification system is more efficient than the system using complicate signal processing process. Hence, the proposed method can be used at analysis and classification for complicated biosignal such as EEC and is expected to give considerable assistance to clinical diagnosis.