Broomrocket: Open Source Text-to-3D Algorithm for 3D Object Placement

Author:

Bonic Sanja1ORCID,Bonic Janos2ORCID,Schmid Stefan1ORCID

Affiliation:

1. TU Berlin, Berlin, Germany

2. FernUniversität Hagen, Hagen, Germany

Abstract

Story writers and other creative professionals often rely on concept artists to visualize and then further iterate on their work during game development and other visualization processes. This exchange and its various stages are time consuming, and there is no easy remedy for creating a walkable 3D concept art without involving a 3D artist yet. As a first step, we present Broomrocket, an open source text-to-3D algorithm for 3D concept art. Broomrocket’s contribution is an object relation and placement algorithm that transforms user input describing a 3D scene given in plain English language into actual models placed in a 3D scene. It runs locally using an existing downloaded natural language processing model and does not require third-party services unless a connection to an online 3D model distribution platform is desired. In that case, Broomrocket will search for the keywords from the user’s narrative input and desired license, and place them in the 3D scene, adding each model’s individual license to a license file for further usage.

Publisher

Association for Computing Machinery (ACM)

Reference27 articles.

1. AN ARCHITECTURAL PATTERN FOR X3D-BASED VIRTUAL ENVIRONMENTS - An Object-oriented Approach

2. Olavo Da Rosa Belloc Rodrigo B. D. Ferraz Marcio Calixto Cabral Roseli De Deus Lopes and Marcelo Knörich Zuffo. 2012. Virtual reality procedure training simulators in X3D. In Proceedings of the 17th International Conference on 3D Web Technology (Web3D’12). 153–160. DOI:10.1145/2338714.2338741

3. NLTK

4. Blender Foundation. 2023. Blender Website. Retrieved June 30 2023 from https://blender.org

5. Epic Games. 2023. Unreal Engine 5: Procedural Content Generation Overview. Retrieved November 11 2023 from https://docs.unrealengine.com/5.2/en-US/procedural-content-generation-overview/

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3