Calving Front Delineation In Synthetic Aperture Radar Imagery With Fixed Training Amount

Delineating the calving front of a marine-terminating glacier in synthetic aperture radar (SAR) imagery. This can, for example, be done through Semantic Segmentation. For this task, the CaFFe dataset set (https://paperswithcode.com/dataset/caffe) needs to be used. Moreover, the epoch number needs to be fixed to an amount such that each network sees 150 times the number of pixels in CaFFe's training set.