2023
Scale-MAE: A Scale-Aware Masked Autoencoder for Multiscale Geospatial Representation Learning
ICCV 2023oral
Large, pretrained models are commonly finetuned with imagery that is heavily augmented to mimic different conditions and scales, with the resulting models used for various tasks with imagery from a range of spatial scales. Such models overlook scale-specific information in the data for scale-depende…