Semiparametric shift estimation for alignment of ECG data

We address the problem of curve alignment with a semiparametric framework, that is without any knowledge of the shape. This study stems from a biological issue, in which we are interested in the estimation of the average heart cycle signal, but wish to estimate it without any knowledge of the pulse shape, which may differ from one patient to another. We suggest in this paper an estimator based on a smoothed functional of the periodogram. Results show better performances than the standard method, as well as its robustness to the noise level.