English

RT-Splatting: Joint Reflection-Transmission Modeling with Gaussian Splatting

Computer Vision and Pattern Recognition 2026-05-19 v1

Abstract

3D Gaussian Splatting (3DGS) enables real-time novel view synthesis with high visual quality. However, existing methods struggle with semi-transparent specular surfaces that exhibit both complex reflections and clear transmission, often producing blurry reflections or overly occluded transmission. To address this, we present RT-Splatting, a framework that disentangles each Gaussian's geometric occupancy from its optical opacity. This factorization yields a unified surface-volume scene representation with a single set of Gaussian primitives. Our hybrid renderer interprets this representation both as a surface to capture high-frequency reflections and as a volume to preserve clear transmission. To mitigate the ambiguity in jointly optimizing reflection and transmission, we introduce Specular-Aware Gradient Gating, which suppresses misleading gradients from highly specular regions into the transmission branch, effectively reducing distracting floaters. Experiments on challenging semi-transparent scenes show that RT-Splatting achieves state-of-the-art performance, delivering high-fidelity reflections and clear transmission with real-time rendering. Moreover, our factorization naturally enables flexible scene editing. The project page is available at https://sjj118.github.io/RT-Splatting.

Keywords

Cite

@article{arxiv.2605.18263,
  title  = {RT-Splatting: Joint Reflection-Transmission Modeling with Gaussian Splatting},
  author = {Ji Shi and Xianghua Ying and Bowei Xing and Ruohao Guo and Wenzhen Yue},
  journal= {arXiv preprint arXiv:2605.18263},
  year   = {2026}
}

Comments

CVPR 2026 Highlight, Project Page: https://sjj118.github.io/RT-Splatting/

R2 v1 2026-07-22T07:18:53.536Z