Ontology-Enhanced Structure-Prior and View-Consistent 3D Reconstruction for Semantic Visual Systems
International Journal on Semantic Web and Information Systems · 20.07.2026 · 22 (1) · 1-22
DOI: 10.4018/ijswis.416291 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal on Semantic Web and Information Systems, erschienen am 20.07.2026. Verfasst von Qianli Ma. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 3.0).
Abstract (Originalsprache)
Recovering a complete 3D object from a single 2D image is important for semantic visual information systems, where geometry should support interpretation and interaction. Existing methods emphasize visual fidelity or geometric plausibility but insufficiently model semantics, spatial relations, and part-whole structures. This study proposed Hero-3D, an ontology-enhanced semantic-aware framework for single-image 3D reconstruction. It extracted semantic priors and organized them into object-, relation-, and part-level representations. These priors guided depth inference, geometric completion, and shape regularization. A Hero view provided a stable semantic-structural reference, while ViewToken-conditioned features were generated under shared semantic and geometric constraints. Semantic fusion and view attention integrated image priors, cross-view anchors, and latent 3D features, which were decoded into triplane and implicit neural field representations. Experiments showed that Hero-3D improved reconstruction quality, multiview consistency, and semantic structure preservation.
Zitieren
APA
Ma, Q. (2026). Ontology-Enhanced Structure-Prior and View-Consistent 3D Reconstruction for Semantic Visual Systems. International Journal on Semantic Web and Information Systems, 22(1), 1-22. https://doi.org/10.4018/ijswis.416291
Vancouver
Ma Q. Ontology-Enhanced Structure-Prior and View-Consistent 3D Reconstruction for Semantic Visual Systems. International Journal on Semantic Web and Information Systems. 2026;22(1):1-22. doi:10.4018/ijswis.416291
BibTeX
@article{Ma2026,
author = {Ma, Qianli},
title = {Ontology-Enhanced Structure-Prior and View-Consistent 3D Reconstruction for Semantic Visual Systems},
journal = {International Journal on Semantic Web and Information Systems},
year = {2026},
volume = {22},
number = {1},
pages = {1--22},
doi = {10.4018/ijswis.416291}
}
EUVIMED