Feng Liu, Yaxing Li, Rui Su, Jianping Huang, Lei Bai
Full waveform inversion (FWI) is a high-resolution subsurface imaging technique, but its effectiveness is limited by challenges such as noise contamination, sparse acquisition, and artifacts from multiparameter coupling. To address these limitations, this study develops a deep reparameterized FWI (DR-FWI) framework, in which subsurface parameters are represented by a deep neural network. Instead of directly optimizing the parameters, DR-FWI optimizes the network weights to reconstruct them, thereby embedding network priors and facilitating optimization. To provide guidelines for the design and usage of DR-FWI, we benchmark two initial model embedding strategies: one involves pretraining the network to generate predefined initial models (pretraining-based), and the other directly adds the network outputs to the initial models, along with three representative architectures (UNet, CNN, MLP). Extensive ablation experiments show that combining CNN with pretraining-based initialization significantly enhances inversion accuracy, offering valuable insights into network design. To further understand the mechanism of DR-FWI, spectral bias analysis reveals that the network first captures low-wavenumber features and progressively reconstructs high-wavenumber details. This learning pattern supports adaptive multi-scale inversion and provides a physically interpretable view of the inversion process. Notably, the robustness of DR-FWI is validated under various noise levels and sparse acquisition scenarios, where its strong performance with limited shots and receivers demonstrates reduced reliance on dense observational data. Additionally, a “backbone-branch” structure is proposed to extend DR-FWI to multiparameter inversion, and its efficacy in mitigating cross-parameter interference is validated on a synthetic anomaly model and the Marmousi2 model. These results suggest a promising direction for joint inversion involving multiple parameters or multiphysics.