few-shot learning for seismic facies segmentation

What is it about?

we have devel- oped a few-shot seismic facies segmentation model. Few-shot learning has been designed to learn to perform with very few labelsandwedesignreconstructingmaskedtracesasapretext task for self-supervised learning to obtain a good feature ex- tractor. By these, this model can use all seismic data from different fields, which is different from image data as the tex- ture-based data. With two different seismic data in turn as a meta-training setanda meta-testing set,ourmodelworkswell inone-andfive-shotsettings,whichmeansonlyonelabeland five labels, respectively.

Featured Image