心磁図のＩＣＡを用いた環境磁気ノイズ除去におけるパラメータ決定の検討 Study of Parameter Decision for a Noise Rejection Method Using ICA for a Magnetocardiogram
We studied an environmental magnetic noise rejection method using independent component analysis (ICA) for a magnetocardiogram (MCG). In the ICA, when the SN ratio of the data is reduced, the accuracy of the noise rejection becomes low and it becomes difficult to reach a decision regarding the contraction dimension. In order to solve these problems, we examined the initial value of the transformation matrix W of the ICA. We proposed an initial value decision procedure for the transformation matrix W, and as a result the accuracy of the noise rejection was improved. Verification was carried out for MCGs with various SN ratios. It was shown that this method was effective for improving noise rejection.
日本応用磁気学会誌 34(2), 146-150, 2010