Application of truncated Newton method in multiparameter FWI imaging

Full Waveform Inversion (FWI) imaging is an inversion technique aimed at simultaneously estimating subsurface velocity and reflectivity. One of the main challenges faced in multiparameter inversion is the coupling effects between different physical properties, known as interparameter crosstalk, which can lead to biased and unreliable model updates. Second-order optimization methods are promising in addressing this problem due to the ability of the Hessian matrix to mitigate multiparameter interferences. In this work, we investigate the application of the truncated Newton algorithm in FWI imaging, aiming to use the action of the multiparameter Hessian matrix as an alternative to reduce crosstalk effects in the estimation of high-resolution images for subsurface structures.