Real-time Global Illumination Decomposition of Videos

We propose the first approach for decomposition of a monocular color video into direct and indirect illumination components in real time. retrieve, separate layers, contribution made to scene appearance by reflectance, light sources reflections from various coherent regions one another. Existing techniques that invert global transport require image capture under multiplexed controlled lighting, or only enable single at slow off-line frame rates. In contrast, our works regular videos produces temporally layers real-time At core are several sparsity priors estimation per-pixel based on small set jointly estimated base reflectance colors. The resulting variational problem uses new formulation sparse dense sets non-linear equations we solve efficiently using novel alternating data-parallel optimization strategy. evaluate qualitatively quantitatively, show improvements over state art this field, both quality runtime. addition, demonstrate editing applications with consistent illumination.