Total variation-based dense depth from multicamera array

摘要: Multicamera arrays are increasingly employed in both consumer and industrial applications, various passive techniques documented to estimate depth from such camera arrays. Current estimation methods provide useful estimations of an imaged scene but often impractical due significant computational requirements. This paper presents a framework that generates high-quality continuous map multicamera array/light-field cameras. The proposed utilizes analysis the local epipolar plane image initiate process. estimated is then refined using total variation minimization based on Fenchel–Rockafellar duality. Evaluation this method well-known benchmark indicates performs well terms accuracy when compared with top-ranked baseline algorithm. test dataset includes photorealistic nonphotorealistic scenes. Notably, requirements required achieve equivalent significantly reduced top algorithms. As consequence, suitable for deployment applications.