Text to 3D Scene Generation with Rich Lexical Grounding
This addresses the need for more flexible and accurate text-to-3D scene generation for applications in art, education, and robotics, representing an incremental advance over prior work.
The paper tackles the problem of generating 3D scenes from natural language descriptions by introducing a dataset with annotations and learning to ground textual terms to objects, resulting in improved scene generation over rule-based methods as shown quantitatively.
The ability to map descriptions of scenes to 3D geometric representations has many applications in areas such as art, education, and robotics. However, prior work on the text to 3D scene generation task has used manually specified object categories and language that identifies them. We introduce a dataset of 3D scenes annotated with natural language descriptions and learn from this data how to ground textual descriptions to physical objects. Our method successfully grounds a variety of lexical terms to concrete referents, and we show quantitatively that our method improves 3D scene generation over previous work using purely rule-based methods. We evaluate the fidelity and plausibility of 3D scenes generated with our grounding approach through human judgments. To ease evaluation on this task, we also introduce an automated metric that strongly correlates with human judgments.