高速ウェーブレット変換およびその脳波解析への応用

概要

論文の詳細を見る
In this paper, we examine a fast method for computing the continuous wavelet transform (Fast Wavelet Transform). We propose some methods to improve accuracy and to compute phase information which arises when attempting to use the fast algorithm in the frequency domain. In order to improve accuracy, we show that it is effective to use mother wavelets whose frequencies are lower by 2 octaves than the Nyquist frequency. This technique doesn't have an influence on the computation speed. Furthermore, in order to get a high degree of accuracy in the high frequency area, we show that it is effective to use up sampling by using a L-Spline interpolation. In addition, the effectiveness of our method for phase information was confirmed using a model signal. As an application, the electroencephalogram (EEG) was analyzed by the fast wavelet transform and several knowledges were obtained.
一般社団法人日本機械学会の論文
1999-05-25