An All-in-One Network for Dehazing and Beyond

TL;DR

This paper introduces AOD-Net, an end-to-end CNN model for image dehazing that outperforms previous methods and enhances high-level vision tasks like object detection on hazy images.

Contribution

AOD-Net is a novel, lightweight CNN that directly generates dehazed images without estimating transmission or atmospheric light separately, enabling seamless integration with other models.

Findings

AOD-Net outperforms state-of-the-art dehazing methods in PSNR and SSIM.

Joint training of AOD-Net with Faster R-CNN improves object detection on hazy images.

AOD-Net is lightweight and easily embedded into other deep learning frameworks.

Abstract

This paper proposes an image dehazing model built with a convolutional neural network (CNN), called All-in-One Dehazing Network (AOD-Net). It is designed based on a re-formulated atmospheric scattering model. Instead of estimating the transmission matrix and the atmospheric light separately as most previous models did, AOD-Net directly generates the clean image through a light-weight CNN. Such a novel end-to-end design makes it easy to embed AOD-Net into other deep models, e.g., Faster R-CNN, for improving high-level task performance on hazy images. Experimental results on both synthesized and natural hazy image datasets demonstrate our superior performance than the state-of-the-art in terms of PSNR, SSIM and the subjective visual quality. Furthermore, when concatenating AOD-Net with Faster R-CNN and training the joint pipeline from end to end, we witness a large improvement of the object detection performance on hazy images.