Answers are grounded in this paper and cite their sources.
We develop an algorithm based on a convolutional neural network (CNN) that detects faults directly from 3D seismic amplitude images. Our network is trained with human-labeled 2D images sliced from 3D cubes, where each pixel is labeled as either fault or non-fault. After training, the CNN predicts fault probabilities at every location in the image. The test results on synthetic and field data images suggest that the CNN predicts fault locations with high accuracy. With the capability of emulating human experience and evolving via training with new datasets, CNN technology shows great promise for automating fault detection.
No takes yet. Share an insight, caveat, or question.
Guo et al. (2018) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: