SiameseFWI: A Deep Learning Network for Enhanced Full Waveform Inversion

• Published in: Journal of geophysical research. Machine learning and computation Vol. 1; no. 3
• Main Authors: Saad, Omar M., Harsuko, Randy, Alkhalifah, Tariq
• Format: Journal Article
• Language: English
• Published: 01.09.2024
• Subjects: full wave inversion; Siamese network
• ISSN: 2993-5210; 2993-5210
• DOI: 10.1029/2024JH000227

The performance of full‐wave inversion (FWI) depends highly on how we compare the simulated data to observed ones. The simplified assumptions used to generate the simulated data make such comparison even harder. To address this challenge, we introduce SiameseFWI, a novel approach to FWI that plays a critical role in the comparative analysis of simulated and observed seismic data. Employing a Siamese network, this methodology transforms the data into a shared latent space, enabling a robust and effective comparison of data representations. SiameseFWI leverages two identical Convolutional Neural Networks with shared weights trained in a self‐supervised framework, eliminating the necessity for labeled data. In each FWI iteration, the Siamese network and the velocity model are updated to minimize Euclidean distance loss between the latent representations of the data. Empirical evaluation conducted on the Marmousi2 and Overthrust models affirms the robust inversion performance of SiameseFWI compared to traditional FWI methodologies. Furthermore, its application to field data from Western Australia demonstrates its strength and efficacy in inversion. Notably, SiameseFWI exhibits robust inversion performance even in the presence of noise or when employing a linear initial model. Plain Language Summary Full‐wave inversion (FWI) aims to reveal subsurface structures by comparing the simulated and observed seismic data. However, this comparison is often complicated by the inherent differences between the naturally noise‐prone field data and data simulated using our computational devices based on restrictive cost‐effective assumptions (such as acoustic media). SiameseFWI, a novel approach, employs a Siamese network to transform data into a latent space for improved comparison. Rigorous evaluations of various data sets, including synthetic and field data, demonstrate its superior performance in inversion tasks. Importantly, despite SiameseFWI is self‐supervised, it introduces minimal computational overhead compared to traditional FWI methods. This SiameseFWI innovation holds promise for enhancing our understanding of subsurface properties. Key Points SiameseFWI, a deep learning approach, addresses the challenge of comparing simulated and observed seismic data in full‐wave inversion (FWI) by transforming data into a shared latent space SiameseFWI introduces a self‐supervised framework, leveraging the Siamese network, with a minor computational overhead compared to traditional FWI methods SiameseFWI demonstrates robust performance with synthetic and field data, highlighting its effectiveness in practical applications