Modeling Dataset-Level Priors With Learnable Probability Tables for Pansharpening (opens in new tab)

Pansharpening is a crucial technique for enhancing the resolution of multispectral images by fusing high-resolution panchromatic (PAN) images with low-resolution multispectral (LRMS) images. While recent methods, such as convolutional neural networks (CNNs) and transformers, have made progress by leveraging sample-level local and global information, they overlook the statistical structure shared across the entire dataset, which we refer to as the dataset-level priors. In this work, we propose...