Efficiently Modeling 3D Scenes from a Single Image

摘要

A proposed system lets users create a 3D scene easily and quickly from a single image. The scene model consists of background and foreground objects whose coordinates the system calculates on the basis of a boundary between the ground plane and a wall plane. The system quickly extracts foreground objects by combining image segmentation and graph-cut-based optimization. It enables efficient modeling of foreground objects, easy creation of their textures, and rapid construction of scene models that are simple but produce sufficient 3D effects.