경계 중심 비전 사전학습으로 밀집 공간 인식 강화
Vision Pretraining for Dense Spatial Perception
- 매체
- Hugging Face
- 게시
- 중요도
- 3/5
- 화제 지수
- 15
한국어 요약
밀집 공간 인식을 위해 경계와 형태 불연속성에 초점을 맞춘 비전 사전학습 방법을 제안했다. 마스크된 경계 모델링 자기지도학습 패러다임을 통해 LingBot-Vision을 개발했으며, 깊이 추정 등 다양한 다운스트림 비전 작업에서 DINOv3 대비 우수한 성능을 입증했다.
원문 발췌 · English
Dense spatial perception is essential for physical intelligence, where visual systems are expected to recover structured, metric, and actionable representations from pixel observations. Modern visual foundation models tend 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 propose masked 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 facilitate dense visual token learning. By scaling this framework, we develop LingBot-Vision and demonstrate its efficacy across a diverse set of downstream vision tasks with DINOv3 as a strong baseline. Remarkably, LingBot-Vision drives the progression from LingBot-Depth 1.0 to LingBot-Depth 2.0 for depth completion, and thereby yields enhanced depth estimation, a key pillar for embodied artificial intelligence. Our findings reveal that boundary modeling goes beyond simple line segments and instead serves as a scalable pretraining principle for learning spatially structured visual representations.