• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.46 no.1
• /
• pp.28-33
• /
• 2009

In the research of speech recognition, locating the beginning and end of a speech utterance in a background of noise is of great importance. The conventional methods for speech endpoint detection are based on two simple time-domain measurements-short-time energy, and short-time zero-crossing rate, which couldn't guarantee the precise results if in the low signal-to-noise ratio environments. This paper proposes a novel approach that finds the Lyapunov exponent of time-domain waveform. This proposed method has no use for obtaining the frequency-domain parameters for endpoint detection process, e.g. Mel-Scale Features, which have been introduced in other paper. Accordingly, this algorithm is low complexity and suitable for Digital Isolated Word Recognition System.

• Journal of Korean Medical Ki-Gong Academy
• /
• v.10 no.1
• /
• pp.130-157
• /
• 2007

The purpose of this research is to seek for efficient health maintenance device and to suggest desirable daily habit, based on the inquiry on interrelationship between workers' daily lives and their heartbeat change level. The paper survey about general features, case history and daily habits was conducted on workers during medical examination in Jeollabukdo, and examined their change of heartbeat as well. The results of research deducted from data analysis are as follows; 1. There found very positive interrelations between time-domain analysis and frequency-domain analysis, and MHR and LF/HF ratio had negative connection with other analyses. 2. The recipient showed high time-domain analysis when they are younger, have worked shorter or have spouse, and it contributes to stable sympathetic nerve and parasympathetic nerve as it stimulates autonomic nervous system. 3. According to the result of frequency-domain analysis, recipients showed higher TP and LF when they are younger, and the highest HF when they are under 34.The level of VLF was higher for university graduates than the ones who finished high school. The recipients showed higher TP and HF when they don't have spouse, and lower TP, LF and HF when they have worked longer. 4. The level of RMSSD and TSRD was high for the people who don't have case history, and HF was high when they don't have any disease in progress. 5. According to the result concerning correlation of daily habits with time-domain analysis and frequency-domain analysis, cigarette, alcohol and sleeping hours don't affect heartbeat change, but the ones who regularly workout showed higher result in every analysis. It shows that the autonomic nervous system of recipients who regularly exercise response more actively. The result mentioned above suggests that the change of heartbeat is a direct index which shows the change of autonomic nervous system, and it depends on the exercise the most. Thus, workout is proved to be the best method in order for workers to take care of their health.

• The Transactions of the Korea Information Processing Society
• /
• v.2 no.5
• /
• pp.691-700
• /
• 1995

Most fractal image codings using fractal concept require long encoding time because a large amount of computation is needed to find an optimal affine transformation point. Such a problem can be solved by designing a block classifier fitted to characteristics of image blocks. In general, it is possible to predict more precise and various types of blocks in frequency domain than in spatial domain. In this paper, we propose a block classifier to predict the block type using characteristics of DCT(Discrete Cosine Transform). This classifier has merits to enhance the quality of decoded images as well as to reduce the encoding time meeting fractal features. AC coefficient values in frequency domain make it possible to predict various types of blocks. As the results, the number of comparisons between a range block and the correspoding domain blocks to reach an optimal affine transformation point can be reduced. Specially, signs of DCT coefficients help to find the optimal affine transformation point with only two isometric transformations by eliminating unnecessary isometric transformations among eight isometric transformations used in traditional fractal codings.

• Proceedings of the Korean Society For Composite Materials Conference
• /
• 2002.05a
• /
• pp.39-43
• /
• 2002

As the concept of the smart structure, monitoring of acoustic emission (AE) can be applied to inspect the fracture of the entire structure in operating condition using built-in sensors. The objective of this study is to find the characteristics of matrix crack signals in composites due to the different specimen shapes. To detect matrix crack signals, we performed tensile tests by changing the thickness, width and length of the specimen. For the quantitative evaluation, time frequency analysis such as short-time Fourier transform (STFT) was used to characterize the matrix crack signals from PZT sensor. The experimental result shows the distinctive signal features in frequency domain due to the different specimen shapes.

• Journal of Biomedical Engineering Research
• /
• v.27 no.5
• /
• pp.237-244
• /
• 2006

This study proposes an approach to the performance improvement of EMG(Electromyogram) pattern recognition. MFCC(Mel-Frequency Cepstral Coefficients)'s approach is molded after the characteristics of the human hearing organ. While it supplies the most typical feature in frequency domain, it should be reorganized to detect the features in EMG signal. And the dynamic aspects of EMG are important for a task, such as a continuous prosthetic control or various time length EMG signal recognition, which have not been successfully mastered by the most approaches. Thus, this paper proposes reorganized MFCC and HMM-GMM, which is adaptable for the dynamic features of the signal. Moreover, it requires an analysis on the most suitable system setting fur EMG pattern recognition. To meet the requirement, this study balanced the recognition-rate against the error-rates produced by the various settings when loaming based on the EMG data for each motion.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.2
• /
• pp.41-47
• /
• 2024

When there were disparities in performance between models trained in the time and frequency domains, even after conducting an ensemble, we observed that the performance of the ensemble was compromised due to imbalances in the individual model performances. Therefore, this paper proposes a leakage detection technique to enhance the accuracy of pipeline leakage detection through a step-wise learning approach that extracts features from both the time and frequency domains and integrates them. This method involves a two-step learning process. In the Stage 1, independent model training is conducted in the time and frequency domains to effectively extract crucial features from the provided data in each domain. In Stage 2, the pre-trained models were utilized by removing their respective classifiers. Subsequently, the features from both domains were fused, and a new classifier was added for retraining. The proposed transfer learning-based feature fusion technique in this paper performs model training by integrating features extracted from the time and frequency domains. This integration exploits the complementary nature of features from both domains, allowing the model to leverage diverse information. As a result, it achieved a high accuracy of 99.88%, demonstrating outstanding performance in pipeline leakage detection.

• Journal of Multimedia Information System
• /
• v.3 no.2
• /
• pp.1-8
• /
• 2016

In this paper, we propose a method to automatically diagnose various diseases. The input data consists of electrocardiograph (ECG) recordings. We extract R-to-R interval (RRI) signals from ECG recordings, which are preprocessed to remove trends and ectopic beats, and to keep the signal stationary. After that, we perform some prospective analysis to extract time-domain parameters, frequency-domain parameters, and nonlinear parameters of the signal. Those parameters are unique for each disease and can be used as the statistical symptoms for each disease. Then, we perform feature selection to improve the performance of the diagnosis classifier. We utilize the selected features to diagnose various diseases using machine learning. We subsequently measure the performance of the machine learning classifier to make sure that it will not misdiagnose the diseases. The first two steps, which are R-to-R extraction and preprocessing, have been successfully implemented with satisfactory results.

• Journal of Electrical Engineering and Technology
• /
• v.9 no.2
• /
• pp.670-676
• /
• 2014

The objective of this research is to utilize the ultrasonic method to analyze the property of partial discharge (PD) which is generated by the winding of the insulation stator in the generator. Therefore, a PD measurement system is built based on ultrasonic and virtual instruments. Three types of PD models (internal PD model, surface PD model and slot PD model) have been constructed. With the analysis of these experimental results, this research has identified the ultrasonic signals of the discharges which were produced by three types of PD models. This analysis shows the different features among these PD types. Both the time domain and frequency domain of the ultrasonic signals are obviously different. In addition, an experiment based on a large rotating machine has been done to analyze ultrasonic noises. The result indicates that the ultrasonic noises can be wiped off by the filters and algorithms. The application of this system is convenient for the detection of early signs of insulation failure, which is an effective method for diagnosis of insulation faults.

• Journal of Electrical Engineering and Technology
• /
• v.2 no.3
• /
• pp.353-357
• /
• 2007

The bearing diagnostics method is presented in this paper using fuzzy inference based on vibration data. Both time-domain and frequency-domain features are used as input data for bearing fault detection. The Adaptive Network based Fuzzy Inference System (ANFIS) and Genetic Algorithm (GA) have been proposed to select the fuzzy model input and output parameters. Training results give the optimized fuzzy inference system for bearing diagnosis based on measured vibration data. The result is also tested with other sets of bearing data to illustrate the reliability of the chosen model.

• The Journal of the Acoustical Society of Korea
• /
• v.36 no.2
• /
• pp.123-129
• /
• 2017

In this paper, the realization of a directional digital hearing aid was considered. Conventional time domain time delay method was replaced with digital filters in order to make any general-purposed DSP (Digital Signal Processing) chip to produce the similar directivity pattern. Both the time delay algorithm and the digital filter algorithm were initially evaluated by Matlab (Matrix laboratory) for comparison, and it was confirmed by CSR 8675 Bluetooth DSP IC (Digital Signal Processing Integrated Circuit) chip firmware realization. Some remote control features by a smart phone was added to the smart hearing aid for user interface easiness.