Learning to Restore Hazy Video: A New Real-World Dataset and A New Method

Most of the existing deep learning-based dehazing methods are trained and evaluated on the image dehazing datasets, where the dehazed images are generated by only exploiting the information from the corresponding hazy ones. On the other hand, video dehazing algorithms, which can acquire more satisfying dehazing results by exploiting the temporal redundancy from neighborhood hazy frames, receive less attention due to the absence of the video dehazing datasets. Therefore, we propose the first REal-world VIdeo DEhazing (REVIDE) dataset which can be used for the supervised learning of the video dehazing algorithms. By utilizing a well-designed video acquisition system, we can capture paired real-world hazy and haze-free videos that are perfectly aligned by recording the same scene (with or without haze) twice. Considering the challenge of exploiting temporal redundancy among the hazy frames, we also develop a Confidence Guided and Improved Deformable Network (CG-IDN) for video dehazing. The experiments demonstrate that the hazy scenes in the REVIDE dataset are more realistic than the synthetic datasets and the proposed algorithm also performs favorably against state-of-the-art dehazing methods.

1. Introduction

Images and videos captured from the hazy scenes inevitably suffer from limited visibility and low color saturation due to the particles in the haze that will scatter and absorption the light and decrease the albedo of the viewed scene. The goal of the dehazing algorithms is to remove the haze and restore a haze-free scene by given a hazy image or video. This problem has received significant attention since the dehazing algorithm is a necessary pre-processing step for many high-level vision tasks (e.g., scene understanding [31] and detection [18]) applied on the outdoor haze, indoor fire, and smoking scenes.