Enhancing 3D Gaussian splatting for low-quality images: semantically guided training and unsupervised quality assessment
The Visual Computer · 2025-06 · 41 (8) · 5815-5833
DOI: 10.1007/s00371-024-03754-z ↗Kurz gefasst
Fachartikel in der Zeitschrift The Visual Computer, online erschienen am 24.12.2024. Verfasst von Zehao Cao und 6 weiteren Autor:innen. Die Arbeit wurde bisher 13-mal zitiert (Crossref) und nennt 44 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
This paper proposes CLIPGaussian, an improved training and rendering method for 3D Gaussian Splatting tailored to address the degradation of reconstruction quality under low-quality image conditions. By integrating a multi-view semantic evaluation module based on CLIP and a 3D spatial consistency loss, CLIPGaussian enhances the model’s sensitivity to object-level semantics, enabling better reconstruction even with limited or blurred training images. Furthermore, we introduce an unsupervised evaluation method leveraging semantic features and explicit 3D Gaussian properties to assess reconstruction quality, providing practical guidance in scenarios where ground truth is unavailable. Experiments on the Mip-NeRF360 dataset demonstrate CLIPGaussian’s superiority, particularly in low-quality image scenarios, validating both the training and evaluation approaches. This work paves the way for more robust and reliable 3D scene reconstruction from imperfect data. Our source code and pre-trained models are available at https://github.com/ZhChessOvO/CLIPGaussian.
Zitieren
APA
Cao, Z., Wang, Z., Zhang, Y., Jin, C., Cai, W., Zeng, Z., & Liu, J. (2025). Enhancing 3D Gaussian splatting for low-quality images: semantically guided training and unsupervised quality assessment. The Visual Computer, 41(8), 5815-5833. https://doi.org/10.1007/s00371-024-03754-z
Vancouver
Cao Z, Wang Z, Zhang Y, Jin C, Cai W, Zeng Z, et al. Enhancing 3D Gaussian splatting for low-quality images: semantically guided training and unsupervised quality assessment. The Visual Computer. 2025;41(8):5815-5833. doi:10.1007/s00371-024-03754-z
BibTeX
@article{Cao2025,
author = {Cao, Zehao and Wang, Zongji and Zhang, Yuanben and Jin, Cheng and Cai, Weinan and Zeng, Zhihong and Liu, Junyi},
title = {Enhancing 3D Gaussian splatting for low-quality images: semantically guided training and unsupervised quality assessment},
journal = {The Visual Computer},
year = {2025},
volume = {41},
number = {8},
pages = {5815--5833},
doi = {10.1007/s00371-024-03754-z}
}
EUVIMED