OptSplat: Recurrent Optimization for Generalizable Reconstruction and Novel View Renderings


Vemburaj Chockalingam Yadav (German Research Center for Artificial Intelligence), Alain Pagani (German Research Center for Artificial Intelligence), Didier Stricker (German Research Center for Artificial Intelligence, RPTU)
The 35th British Machine Vision Conference
PDFPosterVideo (Right click to download)Supplementary

Abstract

We propose an efficient feed-forward model for novel view synthesis and 3D reconstruction based on Gaussian Splatting, featuring a scalable architecture that reliably predicts multi-view depth maps and 3D Gaussian primitives from as few as two input views. Existing multi-view depth estimation techniques typically depend on processing plane-swept cost volumes, which generate probability distributions over a discrete set of candidate depths. This approach limits scalability, especially when finer depth sampling or higher spatial resolution is required. To address this, we design an optimization-inspired architecture $\textit{OptSplat}$, that employs recurrent iterative updates to refine depth maps and pixel-aligned Gaussian primitives based on previous predictions. Our model leverages a unified update operator that iteratively indexes global cost volumes, progressively improving predictions in the joint space of depth and Gaussian parameters. Comprehensive evaluations across the real world datasets of $\textit{RealEstate10K}$, $\textit{ACID}$ and $\textit{DL3DV}$ shows that our model demonstrates strong cross-dataset generalization and competitive rendering quality for novel views compared to the existing works with plane swept cost volumes, while at the same time offering a significant boost in reconstruction and rendering speed, especially for high-resolution inputs.

Citation

@inproceedings{Yadav_2025_BMVC,
author    = {Vemburaj Chockalingam Yadav and Alain Pagani and Didier Stricker},
title     = {OptSplat: Recurrent Optimization for Generalizable Reconstruction and Novel View Renderings},
booktitle = {36th British Machine Vision Conference 2025, {BMVC} 2025, Sheffield, UK, November 24-27, 2025},
publisher = {BMVA},
year      = {2025},
url       = {https://bmva-archive.org.uk/bmvc/2025/assets/papers/Paper_540/paper.pdf}
}


Copyright © 2025 The British Machine Vision Association and Society for Pattern Recognition
The British Machine Vision Conference is organised by The British Machine Vision Association and Society for Pattern Recognition. The Association is a Company limited by guarantee, No.2543446, and a non-profit-making body, registered in England and Wales as Charity No.1002307 (Registered Office: Dept. of Computer Science, Durham University, South Road, Durham, DH1 3LE, UK).

Imprint | Data Protection