• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.1
• /
• pp.30-40
• /
• 2022

Electrocardiogram (ECG) signal's shape and characteristic varies through each individual, so it is difficult to classify with one neural network. It is difficult to classify the given data directly, but if corresponding normal beat is given, it is relatively easy and accurate to classify the beat by comparing two beats. In this study, we classify the ECG signal by generating the reference normal beat through the template cluster, and combining with the input ECG signal. It is possible to detect abnormal beats of various individual's records with one neural network by learning and classifying with the imaged ECG beats which are combined with corresponding reference normal beat. Especially, various neural networks, such as GoogLeNet, ResNet, and DarkNet, showed excellent performance when using the comparative learning. Also, we can confirmed that GoogLeNet has 99.72% sensitivity, which is the highest performance of the three neural networks.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.9
• /
• pp.1840-1846
• /
• 2011

It is necessary to estimate the Doppler spectrum for each range cell for the extraction of useful information from the return echoes in radar systems used for the remote sensing purpose such as detection of moving targets and weather surveillance. The signal amplitude in the beat frequency band is the important parameter in the detection and tracking of targets. However, strong clutter echoes do exist in most radar operation environments and the system phase noise spreads both the clutter and signal echoes of the target. In this paper, the effects of this system phase noise are analyzed concerning the clutter and the signal beat spectrum. It is shown that the separation capability of adjacent beat signal depends on the degree of spread in the clutter and beat signal caused by the radar system phase noise

• Journal of Biomedical Engineering Research
• /
• v.23 no.6
• /
• pp.477-484
• /
• 2002

A pacemaker autosense algorithm with dual thresholds. one for noise or tachyarrhythmia detection (noise threshold, NT) and the other for intrinsic beat detection (sensing threshold. ST), was developed to improve the sensing performance in single pass VDD electrograms. unipolar electrograms, or atrial fibrillation detection. When a deflection in an electrogram exceeds the NT (defined as 50% of 57), the autosense algorithm with dual thresholds checks if the deflection also exceeds the ST. If it does, the autosense algorithm calculates the signal to noise ratio (SNR) of the deflection to the highest deflection detected by NT but lower than ST during the last cardiac cycle. If the SNR 2, the autosense algorithm declares an intrinsic beat detection and calculates the next ST based on the three most recent intrinsic peaks. If the SNR $\geq$2, the autosense algorithm checks the number of deflections detected by NT during the last cardiac cycle in order to determine if it is a noise detection or tachyarrhythmia detection. Usually the autosense algorithm tries to set the 57 at 37.5% of the average of the three intrinsic beats, although it changes the percentage according to event classifications. The autosense algorithm was tested through computer simulation of atrial electrograms from 5 patients obtained during EP study, to simulate a worst sensing situation. The result showed that the ST levels for autosense algorithm tracked the electrogram amplitudes properly, providing more noise immunity whenever necessary. Also, the autosense algorithm with dual thresholds achieved sensing performance as good as the conventional fixed sensitivity method that was optimized retrospectively.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.12
• /
• pp.2511-2517
• /
• 2009

Recently, a FMCW radar is used for the various purposes in the short range detection and tracking of targets. The main advantages of a FMCWradar are the comparative simplicity of implementation and the low peak power transmission characterizing the very low probability of signal interception. Since it uses the frequency modulated continuous wave for transmission and demodulation, the received beat frequency represents the range and Doppler information of targets. Detection and extraction of useful information from targets are performed in this beat frequency domain. Therefore, the resolution and accuracy in the estimation of a beat spectrum are very important. However, using the conventional FFT estimation method, the high resolution spectrum estimation with a low sidelobe level is not possible if the acquisition time is very short in receiving target echoes. This kind of problems deteriorates the detection performance of adjacent targets having the large magnitude differences in return echoes and also degrades the reliability of the extracted information. Therefore, in this paper, the model parameter estimation methods such as autoregressive and eigenvector spectrum estimation are applied to mitigate these problems. Also, simulation results are compared and analyzed for further improvement.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.13 no.1
• /
• pp.1-8
• /
• 2018

This paper proposes a low complexity super resolution algorithm for frequency modulated continuous wave (FMCW) radar systems for foreign object debris (FOD) detection. FOD radar has a requirement to detect foreign object in small units in a large area. However, The fast Fourier transform (FFT) method, which is most widely used in FMCW radar, has a disadvantage in that it can not distinguish between adjacent targets. Super resolution algorithms have a significantly higher resolution compared with the detection algorithm based on FFT. However, in the case of the large number of samples, the computational complexity of the super resolution algorithms is drastically high and thus super resolution algorithms are difficult to apply to real time systems. In order to overcome this disadvantage of super resolution algorithm, first, the proposed algorithm coarsely obtains the frequency of the beat signal by employing FFT. Instead of using all the samples of the beat signal, the number of samples is adjusted according to the frequency of the beat signal. By doing so, the proposed algorithm significantly reduces the computational complexity of multiple signal classifier (MUSIC) algorithm. Simulation results show that the proposed method achieves accurate location even though it has considerably lower complexity than the conventional super resolution algorithms.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.54 no.3
• /
• pp.198-205
• /
• 2005

In this paper, we proposed a new algorithm for the separation of fetal ECG from single channel abdominal ECG. The algorithm consists of a stage of demixing vector calculation for initial signal and a stage of fetal beat detection for the rest of signal. The demixing vector was obtained by applying independent component analysis technique to projected signals into time-frequency domain. For the test of this algorithm, simulation signals, De Lathauwer's data and some measured data, which was acquired from 8 healthy volunteers whose pregnant periods ranged from 22 weeks to 35 weeks and whose ages from 27 to 37, were used. For each data, the accuracy of fetal beat detection was $100\%$ and with the location of fetal beats, fetal heart rate variability and morphology could be offered. In conclusion, this proposed algorithm showed the possibility of fetal beat separation with a single channel abdominal ECG and it might be adopted to a fetal health monitoring system, by which a single channel abdominal ECG is acquired.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.67 no.7
• /
• pp.898-902
• /
• 2018

A novel heart beat interval estimation algorithm is presented based on parabola approximation method. This paper presented a two-step processing scheme; a first stage is finding R-peak in the Electrocardiogram (ECG) by Shannon energy envelope estimator and a secondary stage is computing the interpolated peak location by parabola approximation. Experimental results show that the proposed algorithm performs better than with the previous method using low sampled ECG signals.

• Journal of Biomedical Engineering Research
• /
• v.29 no.6
• /
• pp.466-476
• /
• 2008

The more people use ambulatory electrocardiogram(ECG) for arrhythmia detection, the more researchers report the automatic classification algorithms. Most of the previous studies don't consider the un-balanced data distribution. Even in patients, there are much more normal beats than abnormal beats among the data from 24 hours. To solve this problem, the hierarchical classification using 21 features was adopted for arrhythmia abnormal beat detection. The features include R-R intervals and data to describe the morphology of the wave. To validate the algorithm, 44 non-pacemaker recordings from physionet were used. The hierarchical classification model with 2 stages on domain knowledge was constructed. Using our suggested method, we could improve the performance in abnormal beat classification from the conventional multi-class classification method. In conclusion, the domain knowledge based hierarchical classification is useful to the ECG beat classification with unbalanced data distribution.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.47 no.6
• /
• pp.27-32
• /
• 2010

77GHz FMCW(Frequency Modulation Continuous Wave) radar system has been used for automotive active safety systems. In typical automotive radar, the moving target detection and clutter cancellation including stationary targets are very important signal processing algorithms. This paper proposed the moving target detection algorithm which improve the detection probability and reduce the false alarm rate. First, the proposed moving target beat-frequency extraction filter is used in order to suppress clutter, and then the data association is applied by using the extracted moving target beat-frequency. Then, the zero-Doppler target is eliminated to remove the rest of clutter.

• Journal of Convergence for Information Technology
• /
• v.12 no.4
• /
• pp.260-269
• /
• 2022

In this paper, We propose an anomaly detection model using deep neural network to automate the identification of outliers of the national air pollution measurement network data that is previously performed by experts. We generated training data by analyzing missing values and outliers of weather data provided by the Institute of Environmental Research and based on the BeatGAN model of the unsupervised learning method, we propose a new model by changing the kernel structure, adding the convolutional filter layer and the transposed convolutional filter layer to improve anomaly detection performance. In addition, by utilizing the generative features of the proposed model to implement and apply a retraining algorithm that generates new data and uses it for training, it was confirmed that the proposed model had the highest performance compared to the original BeatGAN models and other unsupervised learning model like Iforest and One Class SVM. Through this study, it was possible to suggest a method to improve the anomaly detection performance of proposed model while avoiding overfitting without additional cost in situations where training data are insufficient due to various factors such as sensor abnormalities and inspections in actual industrial sites.