Full text loading...
-
Data-to-Data and Gradient-to-Gradient Translations in Geophysics Using Deep Neural Networks
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 82nd EAGE Annual Conference & Exhibition, Oct 2021, Volume 2021, p.1 - 5
Abstract
Summary
Image-to-image translation using GANs have successfully been applied to a wide variety of problems, from mundane implementations to turn horses into zebras, to stunning synthetic media deepfakes. We explore its application in geophysics as a cost-reduction tool, and demonstrate its potential in 3D field data. We show it can be used to learn the mapping between different data flavours of interest in modern data processing workflows: acoustic/elastic for full-waveform inversion and geophone vertical-component/hydrophone-pressure for up- and down-going wavefield separation.
© EAGE Publications BV