Vision Pretraining for Dense Spatial Perception
Summary
This paper introduces masked boundary modeling, a self-supervised paradigm for vision pretraining that learns sub-pixel boundary representations to improve dense spatial perception. The resulting model, LingBot-Vision, demonstrates significant improvements in depth estimation and other downstream tasks, showing that boundary modeling is a scalable pretraining principle for spatially structured visual representations.
View Cached Full Text
Cached at: 07/07/26, 06:42 AM
Paper page - Vision Pretraining for Dense Spatial Perception
Source: https://huggingface.co/papers/2607.05247 Published on Jul 6
·
Submitted byhttps://huggingface.co/cherubicxn
Nanon Jul 7
Abstract
Boundary modeling enables dense spatial perception by learning sub-pixel representations that enhance depth estimation and support embodied AI applications.
Densespatial perceptionis essential for physical intelligence, where visual systems are expected to recover structured, metric, and actionable representations from pixel observations. Modernvisual foundation modelstend to prioritize semantic invariance, often at the expense of detailed spatial understanding. In this work, we study vision pretraining through a boundary-centric lens, motivated by the premise that boundaries and shape discontinuities offer essential cues for perceiving geometric properties. Concretely, we proposemasked boundary modeling, a self-supervised paradigm that dynamically learns sub-pixel boundary representations and subsequently leverages the discovered boundary-bearing tokens as masked targets to facilitatedense visual token learning. By scaling this framework, we develop LingBot-Vision and demonstrate its efficacy across a diverse set of downstream vision tasks withDINOv3as a strong baseline. Remarkably, LingBot-Vision drives the progression from LingBot-Depth 1.0 to LingBot-Depth 2.0 fordepth completion, and thereby yields enhanced depth estimation, a key pillar forembodied artificial intelligence. Our findings reveal thatboundary modelinggoes beyond simple line segments and instead serves as a scalable pretraining principle for learning spatially structured visual representations.
View arXiv pageView PDFProject pageGitHub65Add to collection
Get this paper in your agent:
hf papers read 2607\.05247
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2607.05247 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2607.05247 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2607.05247 in a Space README.md to link it from this page.
Collections including this paper1
Similar Articles
LingBot-Vision: masked boundary modeling for self-supervised pretraining (0.296 NYUv2 linear-probe RMSE at 1.1B vs 0.309 for DINOv3-7B, trails on ImageNet); weights in 4 sizes[R]
LingBot-Vision introduces masked boundary modeling for self-supervised pretraining, achieving a 0.296 RMSE on NYUv2 linear-probe with 1.1B parameters versus 0.309 for DINOv3-7B, though it trails on ImageNet; weights are released in four sizes.
@AdinaYakup: LingBot Vision A self-supervised vision backbone family for dense spatial perception from Ant Group @robbyant_brain - A…
LingBot Vision, a self-supervised vision backbone family from Ant Group, uses masked boundary modeling to achieve state-of-the-art performance on dense spatial perception tasks, beating the larger DINOv3 model on NYU-Depth v2.
@ninaddaithankar: Can a vision model learn to see with no augmentations, no masking, no cropping, no reconstruction? It can! Introducing …
Introduces Temporal Difference in Vision (TDV), a novel visual representation learning paradigm that learns useful representations without augmentations, masking, cropping, or reconstruction, and matches state-of-the-art methods on dense spatial tasks.
Beyond 3D VQAs: Injecting 3D Spatial Priors into Vision-Language Models for Enhanced Geometric Reasoning
This paper proposes GASP, a framework that injects geometric priors into vision-language models via deep supervision with contrastive and depth consistency losses, achieving significant improvements on 3D spatial reasoning benchmarks without using 3D VQA data.
Which Pretraining Paradigm Better Serves Spatial Intelligence? An Empirical Comparison of Vision-Language and Video Generation Models
This paper presents a systematic frozen-feature probing study comparing vision-language models (VLMs) and video generation models (VGMs) on spatial intelligence tasks. It finds that VLMs excel at semantic tagging and instance grouping, while VGMs provide better dense geometry and camera motion signals, and a naive fusion of both yields strong performance across all axes.