See Blue Sky: Deep Image Dehaze Using Paired and Unpaired Training Images

TL;DR

This paper introduces a cycle GAN-based model for image dehazing that effectively restores clear blue skies in hazy outdoor images by leveraging both paired and unpaired training data.

Contribution

The paper proposes a novel end-to-end dehazing model using cycle GAN with multiple loss functions to improve sky restoration and image quality.

Findings

Successfully restores clear blue skies in hazy images

Combines paired and unpaired datasets for training

Produces visually appealing dehazed images with minimal distortion

Abstract

The issue of image haze removal has attracted wide attention in recent years. However, most existing haze removal methods cannot restore the scene with clear blue sky, since the color and texture information of the object in the original haze image is insufficient. To remedy this, we propose a cycle generative adversarial network to construct a novel end-to-end image dehaze model. We adopt outdoor image datasets to train our model, which includes a set of real-world unpaired image dataset and a set of paired image dataset to ensure that the generated images are close to the real scene. Based on the cycle structure, our model adds four different kinds of loss function to constrain the effect including adversarial loss, cycle consistency loss, photorealism loss and paired L1 loss. These four constraints can improve the overall quality of such degraded images for better visual appeal and ensure reconstruction of images to keep from distortion. The proposed model could remove the haze of images and also restore the sky of images to be clean and blue (like captured in a sunny weather).