Machine Learning for Visual Recognition in Geosciences

Analyzing images is a big part of day-to-day life of geoscientists, such as conducting seismic interpretation or lithofacies identification and classification. Furthermore, visual representation, recognition and feature extraction play a crucial role in providing a foundation to solve different geosciences research questions, including reconstructing depositional environment, marine ecosystem and tectonic history. Imagine analysis is often costly, time consuming and requires in-depth knowledge of specific geological sub-fields (igneous, metamorphic, sedimentary petrography, and micropaleontology). Recent improvements in machine learning techniques, in particular deep learning, have led to excellent performance in different computer vision tasks (e.g., image classification, segmentation) that significantly increase efficiency and reproducibility. In this course, we will go through the basics of machine learning for visual recognition by analyzing different real-world geoscience problems and try to understand how machine learning algorithms can be used to help solve these problems. This course is intended to provide an introduction to visual recognition with machine learning. No prior knowledge of machine learning and python programming are required.