In recent years, neural-network-based methods for estimating pseudo-well-logs from existing well logs and 3D seismic data have been gaining popularity. Their main advantage over most traditional estimation methods is their ability to extract nonlinear relationships between the seismic data and the sparse set of well logs we want to interpolate. However, the process to go from seismic data and existing well logs to a dense 3D cube of pseudo-well-logs is not simple. The success of the estimation depends on the success of many critical subprocesses and choices that are not trivial and, as far as we know, have not been...

First Page Preview

First page PDF preview
You do not currently have access to this article.