# DensiTok: Making Feed-Forward 3D Gaussian Splatting See More Views Than It Is Given

arXiv: https://arxiv.org/abs/2610.07958 · HTML: https://mlsanity.com/paper/2610.07958/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 47 references that are in the map.
- First posted on arXiv 6 Oct 2026.

## Builds on (most cited first, 11 of 11)

- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [Layer Normalization](https://mlsanity.com/paper/1607.06450/) (2016)
- [Auto-Encoding Variational Bayes](https://mlsanity.com/paper/1312.6114/) (2013)
- [Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow](https://mlsanity.com/paper/2209.03003/) (2022)
- [Depth Anything 3: Recovering the Visual Space from Any Views](https://mlsanity.com/paper/2511.10647/) (2025)
- [Stereo Magnification: Learning View Synthesis using Multiplane Images](https://mlsanity.com/paper/1805.09817/) (2018)
- [Splatt3R: Zero-shot Gaussian Splatting from Uncalibrated Image Pairs](https://mlsanity.com/paper/2408.13912/) (2024)
- [WorldMirror: Universal 3D World Reconstruction with Any-Prior Prompting](https://mlsanity.com/paper/2510.10726/) (2025)
- [VolSplat: Rethinking Feed-Forward 3D Gaussian Splatting with Voxel-Aligned Prediction](https://mlsanity.com/paper/2509.19297/) (2025)
- [GSFixer: Improving 3D Gaussian Splatting with Reference-Guided Video Diffusion Priors](https://mlsanity.com/paper/2508.09667/) (2025)
- [Multi-Fidelity Flow Matching: Cascaded Refinement of PDE Solutions](https://mlsanity.com/paper/2605.16118/) (2026)

## Abstract (arXiv)

Feed-forward 3D Gaussian Splatting (3DGS) reconstructs a scene in a single forward pass, replacing per-scene optimization with a network trained across many scenes. Its quality, however, degrades sharply as the number of input images drops. The bottleneck is upstream of the reconstruction heads: from a few unposed views, the internal representation they read carries no evidence for unobserved regions, leaving holes, floaters, and blur. The common remedy supplies that evidence as pixels, synthesizing extra views with an image or video generator and re-encoding them, which is costly and not 3D-consistent by construction. We instead densify the evidence itself. We present DensiTok, a plug-in module for pretrained feed-forward 3DGS models that densifies their internal geometry tokens directly, making a frozen backbone behave as though it had observed many more views than it was given. DensiTok compresses those tokens into a compact latent space, completes the latents of the unobserved viewpoints in a single flow-matching step conditioned on camera geometry, and decodes them back into tokens that the original reconstruction heads. The same module design can be integrated into different pretrained predictors while keeping each backbone and its reconstruction heads frozen. Completion in a low-dimensional latent space requires no image synthesis or additional encoder passes. Across three pretrained backbones and two benchmarks, DensiTok consistently improves sparse-view reconstruction and recovers much of the gap to dense-view reconstruction.