Information Extraction to Generate Visual Simulations of Car Accidents from Written Descriptions

摘要

This paper describes a system to create animated 3D scenes of car accidents from written reports. The text-to-scene conversion process consists of two stages. An information extraction module creates a tabular description of the accident and a visual simulator generates and animates the scene. We outline the overall structure of the text-to-scene conversion and the template structure. We then describe the information extraction system: the detection of the static objects and the vehicles, their initial directions, the chain of events, and the collisions. We show snapshots of the car animation output and we conclude with the results we obtained.