Pixel-level Scene Understanding in One Token: Visual States Need What-is-Where Composition

TL;DR

CroBo is a novel self-supervised learning framework that encodes detailed scene semantics and spatial configurations into visual states, improving robot policy learning and scene understanding from streaming videos.

Contribution

It introduces CroBo, a global-to-local reconstruction method that explicitly encodes what-is-where in visual states for robotic perception.

Findings

Achieves state-of-the-art results on robot policy benchmarks.

Preserves pixel-level scene composition in learned representations.

Encodes scene dynamics and element interactions effectively.

Abstract

For robotic agents operating in dynamic environments, learning visual state representations from streaming video observations is essential for sequential decision making. Recent self-supervised learning methods have shown strong transferability across vision tasks, but they do not explicitly address what a good visual state should encode. We argue that effective visual states must capture what-is-where by jointly encoding the semantic identities of scene elements and their spatial locations, enabling reliable detection of subtle dynamics across observations. To this end, we propose CroBo, a visual state representation learning framework based on a global-to-local reconstruction objective. Given a reference observation compressed into a compact bottleneck token, CroBo learns to reconstruct heavily masked patches in a local target crop from sparse visible cues, using the global bottleneck token as context. This learning objective encourages the bottleneck token to encode a fine-grained representation of scene-wide semantic entities, including their identities, spatial locations, and configurations. As a result, the learned visual states reveal how scene elements move and interact over time, supporting sequential decision making. We evaluate CroBo on diverse vision-based robot policy learning benchmarks, where it achieves state-of-the-art performance. Reconstruction analyses and perceptual straightness experiments further show that the learned representations preserve pixel-level scene composition and encode what-moves-where across observations. Project page available at: https://seokminlee-chris.github.io/CroBo-ProjectPage.