Structure label prediction using similarity-based retrieval and weakly supervised label mapping

Abstract

Structure label prediction using similaritybased retrieval and weakly supervised label mapping", GEOPHYSICS 2019 84:1, V67-V79. ABSTRACTRecently, there has been significant interest in various supervised machine learning techniques that can help reduce the time and effort consumed by manual interpretation workflows. However, most successful supervised machine learning algorithms require huge amounts of annotated training data. Obtaining these labels for large seismic volumes is a very timeconsuming and laborious task. We address this problem by presenting a weakly-supervised approach for predicting the labels of various seismic structures. By having an interpreter select a very small number of exemplar images for every class of subsurface structures, we use a novel similarity-based retrieval technique to extract thousands of images that contain similar subsurface structures from the seismic volume. By assuming that similar images belong to the same class, we obtain thousands of image-level labels for these images; we validate this assumption in our results section. We then introduce a novel weakly-supervised algorithm for mapping these rough image-level labels into more accurate pixel-level labels that localize the different subsurface structures within the image. This approach dramatically simplifies the process of obtaining labeled data for training supervised machine learning algorithms on seismic interpretation tasks. Using our method we generate thousands of automaticallylabeled images from the Netherlands Offshore F3 block with reasonably accurate pixel-level labels. We believe this work will allow for more advances in machine learning-enabled seismic interpretation. 0

Self-Supervised Delineation of Geological Structures using Orthogonal Latent Space Projection

“…Comparison with a Machine Learning Framework by Alaudah et al (2019a) Our deep learning framework could also be applied to solve a label-mapping problem similar to the problem solved by Alaudah et al (2019a). The author applied a nonnegative matrix factorization (NNMF) algorithm to predict pixel labels from image labels.…”

“…The features learned during factorization were mapped to corresponding images to delineate geological structures. In Figure 13, we attempt to label pixels by mapping image labels learned from our clustering framework to pixel predictions made by our deep learning model and we compare the result with Alaudah et al (2019a)'s framework.…”

“…Salt imaging is challenging due to the rugose structure of salt domes, leading to complex amplitude patterns in salt facies. These complex patterns can conflict with other structural patterns in the feature factorization of Alaudah et al (2019a)'s framework leading to relatively poorer performance. Our method focuses attempts to separate interesting features from the surrounding background, hence it captures complex patterns better than the previous framework.…”

“…(Di et al, 2018b;Wu et al, 2018;Xiong et al, 2018;Araya-Polo et al, 2017;Dramsch and Lüthje, 2018). Several notable works have also attempted to overcome the effect of limited annotated data by employing semi-supervised and weakly supervised techniques (Alaudah and AlRegib, 2016;Alaudah et al, 2019a;Babakhin et al, 2019;Alfarraj and AlRegib, 2019;Wu et al, 2019;Liu et al, 2019;Di et al, 2018c). In these frameworks, there is less dependence on fully labeled data.…”

Time Series Prediction Based on Complex-Valued S-System Model

“…One of the main purposes of time series analysis is to predict the future data based on the existing historical one. Its idea is to search a model or function, in which the past values are set as inputs and the future values are utilized as outputs [1][2][3][4]. Time series exist in almost all fields of natural science and social science, so researches of time series analysis methods are of great significance for prediction, control, and diagnosis of practical issues [5][6][7].…”