Abstract

Haze and smog are among the most common environmental factors impacting imagequality and, therefore, image analysis. This paper proposes an end-to-endgenerative method for image dehazing. It is based on designing a fullyconvolutional neural network to recognize haze structures in input images andrestore clear, haze-free images. The proposed method is agnostic in the sensethat it does not explore the atmosphere scattering model. Somewhatsurprisingly, it achieves superior performance relative to all existingstate-of-the-art methods for image dehazing even on SOTS outdoor images, whichare synthesized using the atmosphere scattering model. Project detail and code can be found here:https://github.com/Seanforfun/GMAN_Net_Haze_Removal

 

Quick Read (beta)