Scene layout in text-to-scene conversion

Abstract

The text-to-scene conversion is such a process that converts the input text into 3D scenes automatically based on the natural language processing. The scene layout is the basic research content and the key section of the text-to-scene conversion. It realizes the automatic layout through identifying the information about scene layout from the input text and relying on a large database of 3D models. In this paper, the method which can extract the information about scene layout from text descriptions has been put forward by using language rules. And on this basis we proposed a method of layout reasoning by combining entity recognition and spatial analysis. The experiments show that the methods can obtain expected results.