Snapshot Multispectral Image Completion Via Self-Dictionary Transformed Tensor Nuclear Norm Minimization With Total Variation

Snapshot multispectral imaging suffers from severely low spatial resolution and degraded signals due to mosaic rearrangement. In order to recover a signal of full bands and full sensor size from a single snapshot, we propose a self-dictionary-transformed tensor nuclear norm and develop a joint optimization with total variation regularization as a convex completion problem. The proposed nuclear nor...