Abstract: Semantic scene completion is a computer vision technique that combines semantic segmentation and shape completion. Its purpose is to infer a complete 3D scene with semantic information from ...