JWCDN embeds a simplified underwater image formation model into a GAN with multi-scale dense encoder-decoder and edge-preserving modules to jointly compensate wavelength attenuation and dehaze images, trained via a novel synthesis method simulating different water types.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.IV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Jointly Adversarial Network to Wavelength Compensation and Dehazing of Underwater Images
JWCDN embeds a simplified underwater image formation model into a GAN with multi-scale dense encoder-decoder and edge-preserving modules to jointly compensate wavelength attenuation and dehaze images, trained via a novel synthesis method simulating different water types.