徳島大学 教育・研究者情報データベース(EDB)

1. (英) Zhang Pu / (日) Pu Zhang / (読) ぷ ちょう

   役割	任意
   貢献度	任意
   学籍番号	推奨	****

2. 張 欽宇

   役割	任意
   貢献度	任意
   学籍番号	推奨

3. 芥川 正武([徳島大学.大学院社会産業理工学研究部.理工学域.電気電子系.電気電子システム分野]/[徳島大学.理工学部.理工学科.電気電子システムコース.電気電子システム講座])

   役割	任意
   貢献度	任意
   学籍番号	推奨

4. 木内 陽介

   役割	任意
   貢献度	任意
   学籍番号	推奨

5. 小中 信典

   役割	任意
   貢献度	任意
   学籍番号	推奨

(英) ECG Signal Pre-processing by Lifting Scheme Constructing Multi-resolution Morphological Decomposition

(日) Lifting Schemeを用いた多重解像度形態分解法による心電図の前処理

(英) The target for Electrocardiogram (ECG) signal pre-procession was baseline correction and noise suppression with as low as possible waveform distortion. Morphological operations based pre-processing methods had superiority on reducing waveform distortion and computing complexity. However, linear structure element would cause block effect which was the reason for waveform distortion. Therefore, a novel pre-processing method based on lifting scheme constructing multi-resolution morphological decomposition was proposed in this study. The algorithm adopted cubic spline interpolation for design of predicting and updating operators to effectively reduce the block effect. By simulation to standard test ECG signals and comparison with other pre-processing algorithms based on wavelet transform and multi-resolution morphological decomposition, it was concluded that this algorithm could achieve good performance in mean absolute differences, max difference and root mean square error. The waveform distortion was visually acceptable and the computing complexity was not high.

(日) 心電図波形の低歪みなベースライン除去および雑音除去について検討している．本研究では多重解像度の形態分解法を応用した手法を提案し，有用性について検討している．