Fog removal algorithms: Survey and perceptual evaluation

In this paper, we evaluate three state-of-the-art fog removal algorithms. The goal of the evaluation is to compare which algorithm performs better with regards to human perception; this has not been done in previous works. Two approaches are used to evaluate these algorithms: one is computing the ratio between the gradient of the visible edges in the images before and after fog removal; another one is using a psychophysical method with human observers and a rank order protocol. Using both computing based and psychophysical based methods allows us to investigate whether they lead to comparable results. The experimental results presented in the paper gives a clear comparison outcome between the three tested algorithms.