GeoFusionLRM: Geometry‐Aware Self‐Correction for Consistent 3D Reconstruction
Computer Graphics Forum · 24.03.2026
DOI: 10.1111/cgf.70325 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computer Graphics Forum, online erschienen am 24.03.2026. Verfasst von Ahmet Burak Yildirim und 3 weiteren Autor:innen. Die Arbeit nennt 40 Literaturangaben. Gefördert von: Türkiye Bilimsel ve Teknolojik Araştırma Kurumu.
Abstract (Originalsprache)
Single‐image 3D reconstruction with large reconstruction models (LRMs) has advanced rapidly, yet reconstructions often exhibit geometric inconsistencies and misaligned details that limit fidelity. We introduce GeoFusionLRM, a geometry‐aware self‐correction framework that leverages the model's own normal and depth predictions to refine structural accuracy. Unlike prior approaches that rely solely on features extracted from the input image, GeoFusionLRM feeds back geometric cues through a dedicated transformer and fusion module, enabling the model to correct errors and enforce consistency with the conditioning image. This design improves the alignment between the reconstructed mesh and the input views without additional supervision or external signals. Extensive experiments demonstrate that GeoFusionLRM achieves sharper geometry, more consistent normals, and higher fidelity than state‐of‐the‐art LRM baselines.
Förderung
- Türkiye Bilimsel ve Teknolojik Araştırma Kurumu
Zitieren
APA
Yildirim, A. B., Saygin, T., Ceylan, D., & Dundar, A. (2026). GeoFusionLRM: Geometry‐Aware Self‐Correction for Consistent 3D Reconstruction. Computer Graphics Forum. https://doi.org/10.1111/cgf.70325
Vancouver
Yildirim AB, Saygin T, Ceylan D, Dundar A. GeoFusionLRM: Geometry‐Aware Self‐Correction for Consistent 3D Reconstruction. Computer Graphics Forum. 2026. doi:10.1111/cgf.70325
BibTeX
@article{Yildirim2026,
author = {Yildirim, Ahmet Burak and Saygin, Tuna and Ceylan, Duygu and Dundar, Aysegul},
title = {GeoFusionLRM: Geometry‐Aware Self‐Correction for Consistent 3D Reconstruction},
journal = {Computer Graphics Forum},
year = {2026},
doi = {10.1111/cgf.70325}
}
EUVIMED