Binary Segmentation of Seismic Facies Using Encoder-Decoder Neural Networks
This work addresses seismic interpretation for geologists by reducing reliance on manual segmentation, though it appears incremental as it builds on encoder-decoder networks with hybrid loss functions.
The authors tackled seismic facies segmentation by developing a deep neural network (DNFS) that achieved state-of-the-art results with highly detailed predictions using fewer parameters than existing methods like StNet and U-Net.
The interpretation of seismic data is vital for characterizing sediments' shape in areas of geological study. In seismic interpretation, deep learning becomes useful for reducing the dependence on handcrafted facies segmentation geometry and the time required to study geological areas. This work presents a Deep Neural Network for Facies Segmentation (DNFS) to obtain state-of-the-art results for seismic facies segmentation. DNFS is trained using a combination of cross-entropy and Jaccard loss functions. Our results show that DNFS obtains highly detailed predictions for seismic facies segmentation using fewer parameters than StNet and U-Net.