A latent graphical model with multi-scale auxiliary maps bounds unpaired joint-distribution error and synthesizes realistic noise pairs that improve real-world and cryo-EM denoising.
Exploring efficient asymmetric blind-spots for self-supervised denoising in real-world scenarios
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Unpaired Joint Distribution Modeling via Multi-Scale Image Representations
A latent graphical model with multi-scale auxiliary maps bounds unpaired joint-distribution error and synthesizes realistic noise pairs that improve real-world and cryo-EM denoising.