대한의용생체공학회:학술대회논문집 (Proceedings of the KOSOMBE Conference)
- 대한의용생체공학회 1996년도 추계학술대회
- /
- Pages.118-121
- /
- 1996
퍼지 클러스터링을 이용한 심전도 신호의 라벨링에 관한 연구
A Study on Labeling of ECG Signal using Fuzzy Clustering
- 공인욱 (연세대학교 공과대학 전기공학과) ;
- 이정환 (연세대학교 공과대학 전기공학과) ;
- 이상학 (연세대학교 공과대학 전기공학과) ;
- 최석준 (연세대학교 공과대학 전기공학과) ;
- 이명호 (연세대학교 공과대학 전기공학과)
- Kong, I.W. (Department of Electrical Engineering, College of Engineering, Yonsei University) ;
- Lee, J.W. (Department of Electrical Engineering, College of Engineering, Yonsei University) ;
- Lee, S.H. (Department of Electrical Engineering, College of Engineering, Yonsei University) ;
- Choi, S.J. (Department of Electrical Engineering, College of Engineering, Yonsei University) ;
- Lee, M.H. (Department of Electrical Engineering, College of Engineering, Yonsei University)
- 발행 : 1996.11.15
초록
This paper describes ECG signal labeling based on Fuzzy clustering, which is necessary at automated ECG diagnosis. The NPPA(Non parametric partitioning algorithm) compares the correlations of wave forms, which tends to recognize the same wave forms as different when the wave forms have a little morphological variation. We propose to apply Fuzzy clustering to ECG QRS Complex labeling, which prevents the errors to mistake by using If-then comparision. The process is divided into two parts. The first part is a parameters extraction process from ECG signal, which is composed of filtering, QRS detection by mapping to a phase space by time delay coordinates and generation of characteristic vectors. The second is fuzzy clustering by FCM(Fuzzy c-means), which is composed of a clustering, an assessment of cluster validity and labeling.
키워드