SSGP: Sparse Spatial Guided Propagation for Robust and Generic Interpolation

Interpolation of sparse pixel information towards a dense target resolution\nfinds its application across multiple disciplines in computer vision.\nState-of-the-art interpolation of motion fields applies model-based\ninterpolation that makes use of edge information extracted from the target\nimage. For depth completion, data-driven learning approaches are widespread.\nOur work is inspired by latest trends in depth completion that tackle the\nproblem of dense guidance for sparse information. We extend these ideas and\ncreate a generic cross-domain architecture that can be applied for a multitude\nof interpolation problems like optical flow, scene flow, or depth completion.\nIn our experiments, we show that our proposed concept of Sparse Spatial Guided\nPropagation (SSGP) achieves improvements to robustness, accuracy, or speed\ncompared to specialized algorithms.\n