A new algorithm for fitting a gaussian function riding on the polynomial background

In this letter, an efficient algorithm is presented for fitting a Gaussian signal riding on a polynomial background. It is shown that the nonlinear least-squares fitting can be transformed into a standard linear least-squares fitting. The proposed method has the advantage of not requiring the initial estimates of the parameters, and it significantly reduces the computational cost. Various applications of this method have been successfully applied to real world problems; including the problem of estimating the parameters of characteristic waveforms on the modeling of electrocardiogram (ECG) signals and the problem of robust ECG RS-amplitude estimation.