Image Destriping

Image destriping aims to remove unwanted vertical or horizontal stripes from images, a common artifact in various imaging modalities like hyperspectral imaging, infrared imaging, and light-sheet microscopy. Current research focuses on developing advanced algorithms, including deep learning-based CycleGAN architectures and novel graph neural networks, often incorporating wavelet transforms and low-rank tensor regularization to effectively model and remove stripe patterns while preserving image details. Successful destriping improves image quality and enables more accurate analysis in diverse scientific fields, ranging from remote sensing and medical imaging to materials science.