Seismic Images Segmentation with Deep Learning

<a href="https://embed.stanford.edu/iframe/?url=https%3A%2F%2Fpurl.stanford.edu%2Fqf836dh0076" class="su-underline">Show Content</a>

Abstract/Contents

Abstract: The task of seismic facies classification with machine learning has gained momentum and is currently receiving significant attention from researchers. The capabilities of modern neural networks encourage experiments dedicated to finding optimal ways of applying them to the seismic images segmentation task. It is, however, non-trivial to apply conventional machine learning techniques to seismic data. Obtaining labels when working with seismic data is very expensive; labeled data is scarce overall and knowledge generally is not transferrable from one dataset to another; representing a distribution of an entire dataset with labeled data of limited size is often challenging. Due to the seismic data volumes, inference time for machine learning algorithms can be a point of concern. These and some other aspects of applying machine learning to the seismic images segmentation task and seismic data in general are what we seek to address in this work. The topic is vast and requires a lot of time and resources to be tackled robustly. We, therefore, do not aim to provide solutions but rather highlight some key issues, give a possible direction for further investigations, and speculate on the results of experiments.

Author	Petrov, Sergei
Thesis advisor	Mukerji, Tapan	https://orcid.org/0000-0003-1711-1850 (unverified)
Degree granting institution	Stanford University, Department of Energy Resources Engineering

Subject	Image segmentation
Subject	Seismic interpretation
Subject	Reservoir characterization
Subject	Semi-supervised learning
Genre	Text
Genre	Thesis

DOI	https://doi.org/10.25740/qf836dh0076
Location	https://purl.stanford.edu/qf836dh0076

Use and reproduction: User agrees that, where applicable, content will not be used to identify or to otherwise infringe the privacy or confidentiality rights of individuals. Content distributed via the Stanford Digital Repository may be subject to additional license and use restrictions applied by the depositor.
License: This work is licensed under a Creative Commons Attribution Non Commercial 4.0 International license (CC BY-NC).

Preferred citation: Petrov, S. (2021). Seismic Images Segmentation with Deep Learning. Stanford Digital Repository. Available at https://purl.stanford.edu/qf836dh0076

Master's Theses, Doerr School of Sustainability

Loading usage metrics...