A self-supervised 3D CNN estimates primaries and a surface operator by minimizing the difference between the input wavefield and a wavefield reconstructed through the SRME equation, with no ground-truth labels.
The established practice of SRME methods consists of two parts: first, predicting the multiples, and second, subtracting the predicted multiples from the full wavefield
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
physics.geo-ph 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Physics-Driven Self-Supervised Deep Learning for Free-Surface Multiple Elimination
A self-supervised 3D CNN estimates primaries and a surface operator by minimizing the difference between the input wavefield and a wavefield reconstructed through the SRME equation, with no ground-truth labels.