From Patches to Deep Learning: Combining Self-Similarity and Neural Networks for Sar Image Despeckling

Loïc Denis (1) , Charles-Alban Deledalle (2) , Florence Tupin (3, 4)

1 LHC - Laboratoire Hubert Curien [Saint Etienne]
2 IMB - Institut de Mathématiques de Bordeaux
3 IDS - Département Images, Données, Signal
4 IMAGES - Image, Modélisation, Analyse, GEométrie, Synthèse

Speckle reduction has benefited from the recent progress in image processing, in particular patch-based non-local filtering and deep learning techniques. These two families of methods offer complementary characteristics but have not yet been combined. We explore strategies to make the most of each approach.

Keywords: patches; non-local filtering; deep learning; speckle; SAR; PolSAR

Domains: Engineering Sciences [physics]; Signal and Image processing; Computer Science [cs]; Machine Learning [cs.LG]

Loïc Denis, Charles-Alban Deledalle, Florence Tupin. From Patches to Deep Learning: Combining Self-Similarity and Neural Networks for Sar Image Despeckling. 2019 IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2019), Jul 2019, Yokohama, Japan. pp.5113-5116, ⟨10.1109/IGARSS.2019.8898473⟩. ⟨ujm-03114605⟩