Thursday 15 June 4 pm CET. Register here: https://attendee.gotowebinar.com/register/7055758961598683735
Join us for an exclusive webinar series featuring two back-to-back sessions that will provide a comprehensive overview of training a 2D CNN model architecture and seamlessly importing it into the OpendTect environment.
Webinar 1: "Building and Training Your 2D CNN Model with OpendTect" In this webinar, we will dive deep into the OpendTect Machine Learning Plugin and guide you through the process of building your own model using Python and Jupyter Notebook. Learn how to import your model into OpendTect and explore fault segmentation techniques using transfer learning with U-Net or related models. Witness a live demonstration of extracting 2D seismic datasets from 3D seismic datasets using OpendTect's user-friendly interface. We will also provide a step-by-step tutorial on customizing models for specific datasets using dGB Earth Science's odpy and dgbpy libraries.
Webinar 2: "Application of Trained Models for Fault Prediction and Interpretation" In the second webinar of this series, the main focus is the application of the newly trained model to predict faults on real seismic data, not before seen in training, and fine-tune the model using transfer learning to learn how to better interpret faults in the novel dataset.
We will introduce exciting new features related to transfer learning and the 2D "fault interpreters" workflow. Explore the training of a 2D U-Net VGG19 architecture while leveraging advanced data augmentation techniques in OpendTect. Witness the model's application to real data within the context of transfer learning.
This webinar series presents a comprehensive workflow that encompasses 2D dataset extraction, fault segmentation, transfer learning U-Net models, and seamless integration with the OpendTect UI. By following this workflow, researchers and practitioners can effectively analyze geoscience data, extracting valuable insights for exploration and subsurface characterization purposes.
📅 Dates: 15 and 22 June
⏰ Time: 4 PM CEST
Register now to secure your spot.
Please note that spaces are limited, so early registration is recommended.
We look forward to having you join us for this enlightening webinar series!
The picture shows OpendTect's new Fault Prediction Model: 2DVGG19 The results are shown before 'thinning' which is a seperate easy step in our Faults and Fractures plugin either by 'skeletonization' or 'thinning' option