Looking Into the Water by Unsupervised Learning of the Surface Shape
This work is significant for underwater imaging and remote sensing, offering a method to see through water more clearly by estimating and correcting for surface distortions.
This paper tackles the problem of removing image distortions caused by water surface refractions when looking into water from the air. The authors propose a model with two neural-field networks to predict water surface height and image color, enabling unsupervised training and outperforming the latest unsupervised image restoration approach on both simulated and real data.
We address the problem of looking into the water from the air, where we seek to remove image distortions caused by refractions at the water surface. Our approach is based on modeling the different water surface structures at various points in time, assuming the underlying image is constant. To this end, we propose a model that consists of two neural-field networks. The first network predicts the height of the water surface at each spatial position and time, and the second network predicts the image color at each position. Using both networks, we reconstruct the observed sequence of images and can therefore use unsupervised training. We show that using implicit neural representations with periodic activation functions (SIREN) leads to effective modeling of the surface height spatio-temporal signal and its derivative, as required for image reconstruction. Using both simulated and real data we show that our method outperforms the latest unsupervised image restoration approach. In addition, it provides an estimate of the water surface.