EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Language Conditioned Radiance Fields for Open Vocabulary Object Retrieval in Interactive Museum Scenes and Heritage der Network Noise Conditions

Harriet Murray, Eunbin Ryu, Ren Hoshino

Global Media and Social Sciences Research Journal · 05.06.2026 · 7 (2) · 113-125

DOI: 10.71465/gmssrj203 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Global Media and Social Sciences Research Journal, online erschienen am 05.06.2026. Verfasst von Harriet Murray und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The digitization of cultural heritage has transitioned from passive three-dimensional archiving to interactive, semantically rich virtual environments. While Neural Radiance Fields have revolutionized novel view synthesis by providing photorealistic reconstructions of complex scenes, these representations traditionally lack semantic understanding, limiting their utility in interactive museum contexts where users may wish to search for objects using natural language. This paper presents a comprehensive framework for Language-Conditioned Radiance Fields specifically tailored for interactive museum object retrieval. By distilling vision-language embeddings from large pre-trained models into the volumetric representation of a radiance field, our approach enables open-vocabulary query localization directly within the three-dimensional space. We detail the system architecture, which seamlessly integrates semantic feature rendering with traditional color and density prediction, allowing users to retrieve specific artifacts, architectural details, or thematic elements using unconstrained text prompts. Extensive experimental evaluations are conducted on a newly curated dataset of complex museum display cases, characterized by challenging lighting conditions, occlusions, and densely packed artifacts. Our analysis demonstrates that the proposed language-conditioned framework significantly outperforms traditional metadata-based retrieval systems and two-dimensional semantic projection methods in both accuracy and spatial localization. Furthermore, we discuss the implications of this technology for virtual reality, educational accessibility, and curatorial workflows, offering a paradigm shift in how digital patrons interact with spatial archives.

Zitieren

APA

Murray, H., Ryu, E., & Hoshino, R. (2026). Language Conditioned Radiance Fields for Open Vocabulary Object Retrieval in Interactive Museum Scenes and Heritage der Network Noise Conditions. Global Media and Social Sciences Research Journal, 7(2), 113-125. https://doi.org/10.71465/gmssrj203

Vancouver

Murray H, Ryu E, Hoshino R. Language Conditioned Radiance Fields for Open Vocabulary Object Retrieval in Interactive Museum Scenes and Heritage der Network Noise Conditions. Global Media and Social Sciences Research Journal. 2026;7(2):113-125. doi:10.71465/gmssrj203

BibTeX
@article{Murray2026,
  author = {Murray, Harriet and Ryu, Eunbin and Hoshino, Ren},
  title = {Language Conditioned Radiance Fields for Open Vocabulary Object Retrieval in Interactive Museum Scenes and Heritage der Network Noise Conditions},
  journal = {Global Media and Social Sciences Research Journal},
  year = {2026},
  volume = {7},
  number = {2},
  pages = {113--125},
  doi = {10.71465/gmssrj203}
}
Kostenlos starten →