Look-Before-Move: Narrative-Grounded World Visual Attention in Dynamic 3D Story Worlds

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Jiaming Bian, Bingliang Li, Yuehao Wu, Pichao Wang, Zhi Wang, Hailan Ma, Huadong Mo, Zhenhong Sun
• Venue: arXiv: cs.AI
• Published: 2026-06-25
• Collected: 2026-06-26
• URL: http://arxiv.org/abs/2606.26964v1
• PDF: https://arxiv.org/pdf/2606.26964v1
• DOI: UNKNOWN

Plain-Language Brief

The researchers propose a method called Look-Before-Move for camera planning in dynamic 3D story worlds. The aim is to enable cameras to decide what to observe and how to move, based on narrative intent and physical constraints. The method separates observation specification from motion execution, using a Semantic Observation Contract, Monte Carlo Viewpoint Search, and Semantic Trajectory Grounding.

What the Document Says

• The research focuses on camera planning in dynamic 3D story worlds.
• The proposed method is named Look-Before-Move.
• The method separates observation specification from motion execution.
• A Semantic Observation Contract is used to convert directorial intent into executable visual constraints.
• Monte Carlo Viewpoint Search is performed to find narrative-compliant and geometrically feasible viewpoints.
• Semantic Trajectory Grounding is applied to connect selected viewpoints into continuous, collision-aware, and temporally coherent camera motion.
• A dynamic 3D Story World Benchmark has been constructed for evaluation.

Key Concepts Mentioned

• camera planning
• dynamic 3D story worlds
• Look-Before-Move
• Semantic Observation Contract
• Monte Carlo Viewpoint Search
• Semantic Trajectory Grounding
• narrative intent
• physical constraints
• dynamic 3D Story World Benchmark

Claims Made by the Document

• The proposed method improves subject perception, intent consistency, and trajectory quality over representative baselines.

Evidence Used

• Experiments show that our framework improves subject perception, intent consistency, and trajectory quality over representative baselines.

Methods / Approach

• Semantic Observation Contract
• Monte Carlo Viewpoint Search
• Semantic Trajectory Grounding

Limitations Stated by the Authors

Not stated in source

Limitations Noted by the Agent

Agent interpretation — not from the source.

Unknown. The paper does not discuss potential limitations of the proposed method.

Open Questions

• What are the representative baselines compared to which the proposed method shows improvement?

Notes for Future Q&A

May include interpretation. Labeled where it does.

The research focuses on camera planning in dynamic 3D story worlds and proposes a method called Look-Before-Move. The method is designed to improve subject perception, intent consistency, and trajectory quality over representative baselines. However, the paper does not discuss what these representative baselines are.

Raw Abstract / Excerpt

As embodied AI and world models increasingly operate in dynamic 3D environments, visual perception must move beyond passively interpreting given observations toward actively deciding what to observe. We study this problem through camera planning in dynamic 3D story worlds, where the camera must not only generate smooth motion, but also decide what visual evidence should be acquired before it moves. We formulate this capability as Narrative-Grounded World Visual Attention, where the camera acts as an embodied observer that determines what to observe, how to compose the observation, and how to shift attention over time under narrative intent and physical 3D constraints. To realize this capability, we propose Look-Before-Move, a camera planning framework that separates observation specification from motion execution. It first builds a Semantic Observation Contract to convert directorial intent into executable visual constraints, then performs Monte Carlo Viewpoint Search to find narrative-compliant and geometrically feasible viewpoints, and finally applies Semantic Trajectory Grounding to connect selected viewpoints into continuous, collision-aware, and temporally coherent camera motion. We further construct a dynamic 3D Story World Benchmark based on StoryBlender, covering 50 stories, 457 scenes, and 1585 shots with animated characters, semantic scene configurations, and executable 3D environments. Experiments show that our framework improves subject perception, intent consistency, and trajectory quality over representative baselines, demonstrating the importance of organizing visual attention before generating camera motion.