Unpaired Deep Image Dehazing Using Contrastive Disentanglement Learning

TL;DR

This paper introduces a novel unpaired image dehazing method that uses contrastive disentanglement learning within a CycleGAN framework to effectively separate haze-related factors from clear image features, improving dehazing performance.

Contribution

It proposes a contrastive disentangled dehazing network (CDD-GAN) that leverages negative adversaries and contrastive loss to enhance factor separation in unpaired image dehazing tasks.

Findings

Outperforms existing unpaired dehazing methods on synthetic datasets.

Effectively disentangles haze factors from clear images using contrastive learning.

Demonstrates robustness on real-world hazy images.

Abstract

We offer a practical unpaired learning based image dehazing network from an unpaired set of clear and hazy images. This paper provides a new perspective to treat image dehazing as a two-class separated factor disentanglement task, i.e, the task-relevant factor of clear image reconstruction and the task-irrelevant factor of haze-relevant distribution. To achieve the disentanglement of these two-class factors in deep feature space, contrastive learning is introduced into a CycleGAN framework to learn disentangled representations by guiding the generated images to be associated with latent factors. With such formulation, the proposed contrastive disentangled dehazing method (CDD-GAN) employs negative generators to cooperate with the encoder network to update alternately, so as to produce a queue of challenging negative adversaries. Then these negative adversaries are trained end-to-end together with the backbone representation network to enhance the discriminative information and promote factor disentanglement performance by maximizing the adversarial contrastive loss. During the training, we further show that hard negative examples can suppress the task-irrelevant factors and unpaired clear exemples can enhance the task-relevant factors, in order to better facilitate haze removal and help image restoration. Extensive experiments on both synthetic and real-world datasets demonstrate that our method performs favorably against existing unpaired dehazing baselines.