# RDGSplat: Render-Dedicated Geometry for Novel View Synthesis

arXiv: https://arxiv.org/abs/2610.09173 · HTML: https://mlsanity.com/paper/2610.09173/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 72 references that are in the map.
- First posted on arXiv 6 Oct 2026.

## Builds on (most cited first, 21 of 21)

- [Depth Anything 3: Recovering the Visual Space from Any Views](https://mlsanity.com/paper/2511.10647/) (2025)
- [MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motion](https://mlsanity.com/paper/2410.03825/) (2024)
- [AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Views](https://mlsanity.com/paper/2505.23716/) (2025)
- [LVSM: A Large View Synthesis Model with Minimal 3D Inductive Bias](https://mlsanity.com/paper/2410.17242/) (2024)
- [Splatt3R: Zero-shot Gaussian Splatting from Uncalibrated Image Pairs](https://mlsanity.com/paper/2408.13912/) (2024)
- [Streaming 4D Visual Geometry Transformer](https://mlsanity.com/paper/2507.11539/) (2025)
- [Point3R: Streaming 3D Reconstruction with Explicit Spatial Pointer Memory](https://mlsanity.com/paper/2507.02863/) (2025)
- [PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting](https://mlsanity.com/paper/2410.22128/) (2024)
- [HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worlds](https://mlsanity.com/paper/2604.14268/) (2026)
- [OmniVGGT: Omni-Modality Driven Visual Geometry Grounded Transformer](https://mlsanity.com/paper/2511.10560/) (2025)
- [AMB3R: Accurate Feed-forward Metric-scale 3D Reconstruction with Backend](https://mlsanity.com/paper/2511.20343/) (2025)
- [SPFSplatV2: Efficient Self-Supervised Pose-Free 3D Gaussian Splatting from Sparse Views](https://mlsanity.com/paper/2509.17246/) (2025)
- [Emergent Extreme-View Geometry in 3D Foundation Models](https://mlsanity.com/paper/2511.22686/) (2025)
- [Car-GS: Addressing Reflective and Transparent Surface Challenges in 3D Car Reconstruction](https://mlsanity.com/paper/2501.11020/) (2025)
- [POMA-3D: The Point Map Way to 3D Scene Understanding](https://mlsanity.com/paper/2511.16567/) (2025)
- [2DGS-R: Revisiting the Normal Consistency Regularization in 2D Gaussian Splatting](https://mlsanity.com/paper/2510.16837/) (2025)
- [Gaussian Set Surface Reconstruction through Per-Gaussian Optimization](https://mlsanity.com/paper/2507.18923/) (2025)
- [AdaptSplat: Adapting Vision Foundation Models for Feed-Forward 3D Gaussian Splatting](https://mlsanity.com/paper/2605.10239/) (2026)
- [VG2GT: Voxel-Gaussian Splatting Visual Geometry Grounded Transformer](https://mlsanity.com/paper/2606.01573/) (2026)
- [Geometry Gaussians: Decoupling Appearance and Geometry in Gaussian Splatting](https://mlsanity.com/paper/2606.05124/) (2026)
- [AESplat: Advancing Pose-Free Feed-Forward 3D Gaussian Splatting via Decoupled Appearance Modeling](https://mlsanity.com/paper/2609.36693/) (2026)

## Abstract (arXiv)

3D foundation models enable efficient novel view synthesis by carrying a Gaussian head on the representation they already use for reconstruction. However, the views they render fall short of the geometry they recover, because that geometry is estimated under a metric objective and never scored on how it renders. Recent methods alleviate this by updating the backbone weights, but they thereby discard the metric predictions the model was built for and must be repeated for every new backbone. To this end, we propose RDGSplat, a framework that decodes a second geometry dedicated to rendering from a frozen 3D foundation model, leaving its metric predictions intact. In particular, we devise Render-Dedicated Geometry Decoding, which duplicates the pretrained decoders and optimizes the duplicates under photometric supervision alone. Then, a Target-Pose Conditioned Adapter is introduced to reformulate the representation those decoders read, conditioned on the target camera pose rather than the target image. Extensive experiments show that RDGSplat improves novel view synthesis across three feed-forward backbones on four benchmarks, with every pretrained weight frozen. On RE10K, it raises WM2.0 from 20.918 to 24.266,dB while training 205.5,M added parameters against a frozen 1.4,B backbone, and the depth and pose the same model predicts are unchanged.