Seismic facies analysis with theory-guided CNN

What is it about?

Our research introduces a novel approach to seismic facies analysis. Unlike conventional image segmentation or classification models, we incorporate knowledge from geophysical science to guide our deep learning model. This method enhances the robustness of our model, particularly when trained with limited data.

Featured Image

Why is it important?

Our research introduces the concept of theory-guided thinking, which involves anchoring deep learning models with knowledge from geophysical science. We have developed an end-to-end seismic facies analysis model that integrates seismic parameter classification and seismic facies geological interpretation. This approach significantly improves the accuracy of seismic facies analysis, resolves intra-facies noise issues, and is easy to use and implement.