Single Image Haze Removal using a Generative Adversarial Network

TL;DR

This paper introduces an end-to-end generative adversarial network approach for single image haze removal, replacing traditional transmission estimation with direct image translation to improve efficiency and output quality.

Contribution

It proposes a novel GAN-based model using a Tiramisu generator and patch discriminator, with a hybrid loss for enhanced haze removal from single images.

Findings

Performs competitively with state-of-the-art methods on synthetic and real images

Uses a Tiramisu model for better parameter efficiency and performance

Employs a hybrid weighted loss to improve perceptual quality

Abstract

Traditional methods to remove haze from images rely on estimating a transmission map. When dealing with single images, this becomes an ill-posed problem due to the lack of depth information. In this paper, we propose an end-to-end learning based approach which uses a modified conditional Generative Adversarial Network to directly remove haze from an image. We employ the usage of the Tiramisu model in place of the classic U-Net model as the generator owing to its higher parameter efficiency and performance. Moreover, a patch based discriminator was used to reduce artefacts in the output. To further improve the perceptual quality of the output, a hybrid weighted loss function was designed and used to train the model. Experiments on synthetic and real world hazy images demonstrates that our model performs competitively with the state of the art methods.