Abstract: Image captioning is an emerging field at the intersection of computer vision and natural language processing (NLP). It has shown great potential to enhance accessibility by automatically ...

Abstract: Recently, single-image SVBRDF capture is formulated as a regression problem, which uses a network to infer four SVBRDF maps from a flash-lit image. However, the accuracy is still not ...