{"slug": "regimevggt-layer-wise-spatially-preserving-redundancy-removal-for-visual", "title": "RegimeVGGT: Layer-Wise Spatially Preserving Redundancy Removal for Visual Geometry Grounded Transformer", "summary": "Researchers introduced RegimeVGGT, a training-free accelerator for Visual Geometry Grounded Transformer (VGGT) that achieves a 6.7x speedup over VGGT* at matched reconstruction quality by applying layer-wise U-shaped compression along two axes, including Saliency-Guided Banded Merging and Selectively Protected K/V Downsampling, based on spectral, probing, and causal analyses revealing three distinct regimes in the model's layers.", "body_md": "arXiv:2606.18439v1 Announce Type: new\nAbstract: Visual Geometry Grounded Transformer (VGGT) recovers dense 3D scene structure from multi-view images in one forward pass, but quadratic cross-frame attention limits its scalability. Existing training-free accelerators reduce computation uniformly along one axis, missing layer heterogeneity. Our spectral, probing, and causal analyses reveal three regimes: shallow layers lack cross-view structure, middle layers drive cross-view alignment, and deep layers are redundant for dense geometry yet their cross-frame attention remains essential for pose. RegimeVGGT applies layer-wise U-shaped compression along two axes: Saliency-Guided Banded Merging protects geometry- and edge-salient tokens, while Selectively Protected K/V Downsampling preserves cross-frame spatial coverage and the pose-critical path through a phase-shifted spatial grid, a reference-frame anchor, and uncompressed camera/register tokens. Training-free, RegimeVGGT achieves a 6.7x speedup over VGGT* at matched reconstruction quality.", "url": "https://wpnews.pro/news/regimevggt-layer-wise-spatially-preserving-redundancy-removal-for-visual", "canonical_source": "https://arxiv.org/abs/2606.18439", "published_at": "2026-06-18 04:00:00+00:00", "updated_at": "2026-06-19 02:00:40.046673+00:00", "lang": "en", "topics": ["computer-vision", "machine-learning", "large-language-models"], "entities": ["RegimeVGGT", "Visual Geometry Grounded Transformer", "VGGT"], "alternates": {"html": "https://wpnews.pro/news/regimevggt-layer-wise-spatially-preserving-redundancy-removal-for-visual", "markdown": "https://wpnews.pro/news/regimevggt-layer-wise-spatially-preserving-redundancy-removal-for-visual.md", "text": "https://wpnews.pro/news/regimevggt-layer-wise-spatially-preserving-redundancy-removal-for-visual.txt", "jsonld": "https://wpnews.pro/news/regimevggt-layer-wise-spatially-preserving-redundancy-removal-for-visual.jsonld"}}