# Less Context, Better Geometry: Masked Geometric Encoder for Robust 3D Foundation Models

arXiv: https://arxiv.org/abs/2610.06813 · HTML: https://mlsanity.com/paper/2610.06813/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Token pruning and merging for ViTs, one of 12 directions in Efficient LLM inference and serving.
- Builds on 11 of its 64 references that are in the map.
- First posted on arXiv 5 Oct 2026.

## Placed in

- Efficient LLM inference and serving / [Token pruning and merging for ViTs](https://mlsanity.com/inference/token-pruning-and-merging-for-vits/)

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

- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [LoFTR: Detector-Free Local Feature Matching with Transformers](https://mlsanity.com/paper/2104.00680/) (2021)
- [Depth Anything 3: Recovering the Visual Space from Any Views](https://mlsanity.com/paper/2511.10647/) (2025)
- [ARKitScenes: A Diverse Real-World Dataset For 3D Indoor Scene Understanding Using Mobile RGB-D Data](https://mlsanity.com/paper/2111.08897/) (2021)
- [Virtual KITTI 2](https://mlsanity.com/paper/2001.10773/) (2020)
- [VGGT-Long: Chunk it, Loop it, Align it -- Pushing VGGT's Limits on Kilometer-scale Long RGB Sequences](https://mlsanity.com/paper/2507.16443/) (2025)
- [FastVGGT: Training-Free Acceleration of Visual Geometry Transformer](https://mlsanity.com/paper/2509.02560/) (2025)
- [RoMa v2: Harder Better Faster Denser Feature Matching](https://mlsanity.com/paper/2511.15706/) (2025)
- [VGGT-Omega](https://mlsanity.com/paper/2605.15195/) (2026)
- [STAC: Plug-and-Play Spatio-Temporal Aware Cache Compression for Streaming 3D Reconstruction](https://mlsanity.com/paper/2603.20284/) (2026)

## Abstract (arXiv)

Recent progress in 3D foundation models has enabled rapid 3D reconstruction and camera calibration by leveraging learned 3D priors from vast amount of spatial data. However, the all-to-all global attention design leads to quadratic complexity and limits long-sequence inference; unconstrained cross-view interactions also can propagate unreliable evidence from occluded or visually similar but geometrically distant views. In this paper, We introduce a Masked Geometric Encoder (MGE), which promotes the learning of robust geometric representations under incomplete cross-view context. During training, MGE strategically drops frame tokens from global attention and distills from a pretrained full-context teacher model. This allows the model to learn an intrinsically richer per-frame representation while providing sufficient intermediate supervision to avoid performance degradation. Through extensive experiments, we show that MGE leads to much stronger performance under occlusion and doppelganger views while retaining high performance on standard benchmarks. Such a richer frame representation also leads to more effective token reduction during inference. To this end, we develop a novel Anchor-Guided Adaptive token merging technique that preserves representative anchor frames while jointly merging redundant tokens from the remaining views. Compared to other efficient inference approaches, we can achieve inference speedup while consistently maintaining higher reconstruction quality, particularly in limited-view settings.