Automatic salt-body classification using deep-convolutional neural network

We apply deep learning techniques to the problem of the salt body detection in seismic images. We consider salt body classification as an image segmentation problem, and propose to design a multi-layer convolutional neural network, feed in training data to train this network, and test the model using blind test data. Our results indicate that the proposed network architecture and workflow are capable of capturing subtle salt features automatically without the need for manual input. Trained with a limited amount of inline sections, the model can generalize to the blind test data and be efficiently applied to a whole 3D volume of seismic data. Presentation Date: Tuesday, October 16, 2018 Start Time: 8:30:00 AM Location: 204B (Anaheim Convention Center) Presentation Type: Oral