Iris3D: 3D Generation via Synchronized Diffusion Distillation
ACM Transactions on Graphics · 28.02.2026 · 45 (1) · 1-13
DOI: 10.1145/3759249 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM Transactions on Graphics, online erschienen am 11.09.2025. Verfasst von Yixun Liang und 7 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 71 Literaturangaben.
Abstract (Originalsprache)
We introduce Iris3D, a novel 3D content generation system that generates vivid textures and detailed 3D shapes while preserving the input information. Our system integrates a Multi-View Large Reconstruction Model (MVLRM [Li et al. 2023b ]) to generate a coarse 3D mesh and introduces a novel optimization scheme called Synchronized Diffusion Distillation (SDD) for refinement. Unlike previous refined methods based on Score Distillation Sampling (SDS), which suffer from unstable optimization and geometric over-smoothing due to ambiguities across different views and modalities, our method effectively distills consistent multi-view and multi-modal priors from 2D diffusion models in a training-free manner. This enables robust optimization of 3D representations. Additionally, because SDD is training-free, it preserves the diffusion’s prior knowledge and mitigates potential degradation. This characteristic makes it highly compatible with advanced 2D diffusion techniques like IP-Adapters and ControlNet, allowing for more controllable 3D generation with additional conditioning signals. Experiments demonstrate that our method produces high-quality 3D results with plausible textures and intricate geometric details.
Zitieren
APA
Liang, Y., Li, W., Chen, R., Tian, F., Liu, J., Chen, Y., Tan, P., & Long, X. (2026). Iris3D: 3D Generation via Synchronized Diffusion Distillation. ACM Transactions on Graphics, 45(1), 1-13. https://doi.org/10.1145/3759249
Vancouver
Liang Y, Li W, Chen R, Tian F, Liu J, Chen Y, et al. Iris3D: 3D Generation via Synchronized Diffusion Distillation. ACM Transactions on Graphics. 2026;45(1):1-13. doi:10.1145/3759249
BibTeX
@article{Liang2026,
author = {Liang, Yixun and Li, Weiyu and Chen, Rui and Tian, Fei-Peng and Liu, Jiarui and Chen, Ying-Cong and Tan, Ping and Long, Xiao-Xiao},
title = {Iris3D: 3D Generation via Synchronized Diffusion Distillation},
journal = {ACM Transactions on Graphics},
year = {2026},
volume = {45},
number = {1},
pages = {1--13},
doi = {10.1145/3759249}
}
EUVIMED