Efficient Text-Guided 3D-Aware Generation With Score Distillation on 3D Distribution
IEEE Transactions on Circuits and Systems for Video Technology · 2025-10 · 35 (10) · 9865-9877
DOI: 10.1109/tcsvt.2025.3559931 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Circuits and Systems for Video Technology, 2025-10. Verfasst von Yiji Cheng und 7 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 92 Literaturangaben. Gefördert von: Shenzhen Science and Technology Program, Research Fund of Tsinghua University—Tencent Joint Laboratory for Internet Innovation Technology.
Abstract (Originalsprache)
Text-to-3D generation enables the creation of 3D content with infinite possibilities. Existing methods typically involve training 3D generative models, which suffer from poor semantic alignment due to the scarcity of paired 3D data, or optimizing a 3D representation with 2D diffusion guidance, resulting in slow inference, low diversity, and Janus problems. In this paper, we introduce InstantDreamer, a model designed for text-guided 3D-aware generation in a single forward pass without requiring paired training datasets, thereby enhancing efficiency. To accomplish this, we extend score distillation to learn a 3D-aware semantics distribution. We distill priors from diffusion models into a 3D-aware generator, amortizing the optimization time required for new prompts and eliminating the necessity of paired training data. We equip the generator with hierarchical semantics conditioning, explicitly allowing the model to perceive the correspondence between the text distribution and the 3D latent space. Our elaborate designs empower our 3D generative model with multi-view semantic consistency and feed-forward 3D generation capabilities, thus eliminating the need for score distillation-based optimization for each prompt. Both quantitative and qualitative results on the mainstream benchmarks demonstrate that our InstantDreamer generates competitive multi-view semantic consistent 3D assets compared with state-of-the-art methods. Our method outperforms previous approaches in terms of CLIP R-Precision (66.31) and FID (28.47) while also exhibiting a significant boost in generation speed.
Förderung
- Shenzhen Science and Technology Program
- Research Fund of Tsinghua University—Tencent Joint Laboratory for Internet Innovation Technology
Zitieren
APA
Cheng, Y., Yin, F., Huang, X., Yu, X., Liu, J., Feng, S., Yang, Y., & Tang, Y. (2025). Efficient Text-Guided 3D-Aware Generation With Score Distillation on 3D Distribution. IEEE Transactions on Circuits and Systems for Video Technology, 35(10), 9865-9877. https://doi.org/10.1109/tcsvt.2025.3559931
Vancouver
Cheng Y, Yin F, Huang X, Yu X, Liu J, Feng S, et al. Efficient Text-Guided 3D-Aware Generation With Score Distillation on 3D Distribution. IEEE Transactions on Circuits and Systems for Video Technology. 2025;35(10):9865-9877. doi:10.1109/tcsvt.2025.3559931
BibTeX
@article{Cheng2025,
author = {Cheng, Yiji and Yin, Fei and Huang, Xiaoke and Yu, Xintong and Liu, Jiaxiang and Feng, Shikun and Yang, Yujiu and Tang, Yansong},
title = {Efficient Text-Guided 3D-Aware Generation With Score Distillation on 3D Distribution},
journal = {IEEE Transactions on Circuits and Systems for Video Technology},
year = {2025},
volume = {35},
number = {10},
pages = {9865--9877},
doi = {10.1109/tcsvt.2025.3559931}
}
EUVIMED