Unsupervised Single Image Dehazing Using Dark Channel Prior Loss

Single image dehazing is a critical stage in many modern-day autonomous vision applications. Early prior-based methods often involved a time-consuming minimization of a hand-crafted energy function. Recent learning-based approaches utilize the representational power of deep neural networks (DNNs) to learn the underlying transformation between hazy and clear images. Due to inherent limitations in collecting matching clear and hazy images, these methods resort to training on synthetic data, constructed from indoor images and corresponding depth information. This may result in a possible domain shift when treating outdoor scenes. We propose a completely unsupervised method of training via minimization of the well-known, Dark Channel Prior (DCP) energy function. Instead of feeding the network with synthetic data, we solely use real-world outdoor images and tune the network's parameters by directly minimizing the DCP. Although our “Deep DCP” technique can be regarded as a fast approximator of DCP, it actually improves its results significantly. This suggests an additional regularization obtained via the network and learning process. Experiments show that our method performs on par with large-scale supervised methods.

I. Introduction

Haze is an atmospheric phenomenon where small particles, called aerosols, obstruct the clarity of an outdoor scene and lead to poor contrast and loss of detail. The existence of haze affects an image in two aspects. It attenuates the scene radiance with correspondence to an object’s distance from the camera. Moreover, it introduces an additional ambient light component, called the airlight, which causes a “veiling effect” over the clear image. The formation of a hazy image is often described as a linear per-pixel combination of the clear scene radiance and the airlight; the effect of each component is controlled by the transmission map. To recover the scene radiance image, one has to solve a system of linear equations with unknowns (where is the number of image pixels).