원문정보
보안공학연구지원센터(IJSIP)
International Journal of Signal Processing, Image Processing and Pattern Recognition
Vol.7 No.2
2014.04
pp.211-222
피인용수 : 0건 (자료제공 : 네이버학술정보)
초록
영어
The aim of this paper is twofold. First, we define an ECG feature parameter set (32 features) which could represent ECG signal as adequately as possible for diagnosing requirements. Second, we design an automatic classification framework. After benchmark point detection, feature parameter will be extracted. And then the classifier methods and its comparison based on SVM and QNN are presented. The long-term objective is to design a thorough system to realize the recognition of real-time ECG signal and enhance medical treatment.
목차
Abstract
1. Introduction
2. Feature Extraction and Selection
2.1. Feature Extraction using Wavelet Transform
2.2. Feature Selection Methods
3. Classifier Design
3.1. SVM Classifier
3.2. QNN Classifier
4. Experiment and Comparison
4.1. Experiment result with BP and RBF Neural Network Classifiers
4.2. Experiment Result with SVM Classifier
4.3. Experiment Result with QNN Classifier
5. Conclusion and Expectation
Acknowledgements
References
1. Introduction
2. Feature Extraction and Selection
2.1. Feature Extraction using Wavelet Transform
2.2. Feature Selection Methods
3. Classifier Design
3.1. SVM Classifier
3.2. QNN Classifier
4. Experiment and Comparison
4.1. Experiment result with BP and RBF Neural Network Classifiers
4.2. Experiment Result with SVM Classifier
4.3. Experiment Result with QNN Classifier
5. Conclusion and Expectation
Acknowledgements
References
키워드
저자정보
참고문헌
자료제공 : 네이버학술정보