SphereDRUNet: A Spherical Denoiser for Omnidirectional Images
Rita Fermanian, Thomas Maugey, Christine M. Guillemot · 2023
Image denoising is a primary pre-processing task in image processing. Although it has garnered significant research attention in the context of traditional 2D images, omnidirectional image denoising has received relatively limited attention in the literature. Furthermore, extending processing models and tools designed for 2D images to the sphere presents many challenges due to the inherent distortions and non-uniform pixel distributions associated with spherical representations and their underlying projections. In this paper, we address the problem of omnidirectional image denoising and we aim to study the advantage of denoising the spherical image directly rather than its mapping. We introduce a novel network called SphereDRUNet to denoise spherical images using deep learning tools on a spherical sampling. We show that denoising directly the sphere using our network gives better performance, compared to denoising the projected equirectangular images with a similarly learned model.