FlowR: Flowing from Sparse to Dense 3D Reconstructions
Tobias Fischer, Samuel Rota Bulò, Yung-Hsu Yang, Nikhil Keetha, Lorenzo Porzi, Norman Müller, Katja Schwarz, Jonathon Luiten, Marc Pollefeys, Peter Kontschieder · 2025
Read FlowR: Flowing from Sparse to Dense 3D Reconstructions by Tobias Fischer, Samuel Rota Bulò, Yung-Hsu Yang, Nikhil Keetha, Lorenzo Porzi, Norman Müller, Katja Schwarz, Jonathon Luiten, Marc Pollefeys, Peter Kontschieder (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-01647.txt.
Description
Tobias Fischer and a large team teach a generative model to imagine the extra viewpoints that a sparse three-dimensional scan is missing. Gaussian splatting renders scenes beautifully in real time, but quality collapses away from the photographed angles, and dense photo shoots are slow and costly. Earlier fixes distilled two-dimensional generators or hallucinated extra views from a few references, producing inconsistent inventions that corrupted the reconstruction. Writing in twenty twenty-five, the team's flow matching model instead learns the direct path from a sparse reconstruction's renderings to the renderings a dense capture would give.
Trained on millions of paired images, the model processes dozens of views in one pass on a single high-end graphics card. Pipelines using it beat earlier works on standard view-synthesis benchmarks in both sparse and dense regimes, the authors report. Virtual reality capture is the motivating customer.
Catalog record
| CREATOR | Tobias Fischer, Samuel Rota Bulò, Yung-Hsu Yang, Nikhil Keetha, Lorenzo Porzi, Norman Müller, Katja Schwarz, Jonathon Luiten, Marc Pollefeys, Peter Kontschieder |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-01647