Rapid seismic domain transfer: Seismic velocity inversion and modeling using deep generative neural networks

TL;DR

This paper introduces a deep generative neural network approach for rapid seismic velocity inversion, significantly reducing computational costs compared to traditional physics-based methods by framing the problem as domain transfer.

Contribution

The authors develop a deep convolutional GAN that efficiently generates realistic seismic velocity models from data, bypassing complex traditional inversion techniques.

Findings

The neural network produces realistic velocity models from real seismic data.

The approach is robust against velocity errors and artifacts.

Training on synthetic data generalizes well to real datasets.

Abstract

Traditional physics-based approaches to infer sub-surface properties such as full-waveform inversion or reflectivity inversion are time-consuming and computationally expensive. We present a deep-learning technique that eliminates the need for these computationally complex methods by posing the problem as one of domain transfer. Our solution is based on a deep convolutional generative adversarial network and dramatically reduces computation time. Training based on two different types of synthetic data produced a neural network that generates realistic velocity models when applied to a real dataset. The system's ability to generalize means it is robust against the inherent occurrence of velocity errors and artifacts in both training and test datasets.