Dual-Branch Single-Image Self-Supervised Denoising Network Based on Laplacian Nonlocal Self-Similarity Sampling (opens in new tab)

Optical remote sensing images are widely used in numerous fields, and high-quality images generally play a critical role in downstream tasks. In recent years, self-supervised image denoising methods based on deep learning have achieved significant progress. However, existing sampling strategies in self-supervised approaches are susceptible to interference from high-level noise signals, and the network architectures often prioritize pixel-level image reconstruction while lacking attention to h...