# Out of Sight, Not Out of Context? Egocentric Spatial Reasoning in VLMs Across Disjoint Frames

arXiv: https://arxiv.org/abs/2505.24257 · HTML: https://mlsanity.com/paper/2505.24257/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 20 of its 27 references that are in the map.
- First posted on arXiv 30 May 2025.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)

## Builds on (most cited first, 20 of 20)

- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [Ego4D: Around the World in 3,000 Hours of Egocentric Video](https://mlsanity.com/paper/2110.07058/) (2021)
- [MVBench: A Comprehensive Multi-modal Video Understanding Benchmark](https://mlsanity.com/paper/2311.17005/) (2023)
- [Scaling Egocentric Vision: The EPIC-KITCHENS Dataset](https://mlsanity.com/paper/1804.02748/) (2018)
- [SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities](https://mlsanity.com/paper/2401.12168/) (2024)
- [EgoSchema: A Diagnostic Benchmark for Very Long-form Video Language Understanding](https://mlsanity.com/paper/2308.09126/) (2023)
- [Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces](https://mlsanity.com/paper/2412.14171/) (2024)
- [TempCompass: Do Video LLMs Really Understand Videos?](https://mlsanity.com/paper/2403.00476/) (2024)
- [LongVILA: Scaling Long-Context Visual Language Models for Long Videos](https://mlsanity.com/paper/2408.10188/) (2024)
- [What's "up" with vision-language models? Investigating their struggle with spatial reasoning](https://mlsanity.com/paper/2310.19785/) (2023)
- [The EPIC-KITCHENS Dataset: Collection, Challenges and Baselines](https://mlsanity.com/paper/2005.00343/) (2020)
- [LLaVA-3D: A Simple yet Effective Pathway to Empowering LMMs with 3D-awareness](https://mlsanity.com/paper/2409.18125/) (2024)
- [Learning to Track with Object Permanence](https://mlsanity.com/paper/2103.14258/) (2021)
- [Improved Visual-Spatial Reasoning via R1-Zero-Like Training](https://mlsanity.com/paper/2504.00883/) (2025)
- [V-IRL: Grounding Virtual Intelligence in Real Life](https://mlsanity.com/paper/2402.03310/) (2024)
- [Scaling and Beyond: Advancing Spatial Reasoning in MLLMs Requires New Recipes](https://mlsanity.com/paper/2504.15037/) (2025)
- [Black Swan: Abductive and Defeasible Video Reasoning in Unpredictable Events](https://mlsanity.com/paper/2412.05725/) (2024)
- [EgoSpeak: Learning When to Speak for Egocentric Conversational Agents in the Wild](https://mlsanity.com/paper/2502.14892/) (2025)
- [Embodied Executable Policy Learning with Language-based Scene Summarization](https://mlsanity.com/paper/2306.05696/) (2023)

## Built on by (most cited first, 6 of 6)

- [Multimodal Spatial Reasoning in the Large Model Era: A Survey and Benchmarks](https://mlsanity.com/paper/2510.25760/) (2025)
- [More Images, More Problems? A Controlled Analysis of VLM Failure Modes](https://mlsanity.com/paper/2601.07812/) (2026)
- [Mind Meets Space: Rethinking Agentic Spatial Intelligence from a Neuroscience-inspired Perspective](https://mlsanity.com/paper/2509.09154/) (2025)
- [SpaMEM: Benchmarking Dynamic Spatial Reasoning via Perception-Memory Integration in Embodied Environments](https://mlsanity.com/paper/2604.22409/) (2026)
- [OVO-S-Bench: A Hierarchical Benchmark for Streaming Spatial Intelligence in Multimodal LLMs](https://mlsanity.com/paper/2606.03890/) (2026)
- [Uruqi: Learning Spatial Cognition from Visual Experience](https://mlsanity.com/paper/2609.39195/) (2026)

## Abstract (arXiv)

An embodied AI assistant operating on egocentric video must integrate spatial cues across time - for instance, determining where an object A, glimpsed a few moments ago lies relative to an object B encountered later. We introduce Disjoint-3DQA , a generative QA benchmark that evaluates this ability of VLMs by posing questions about object pairs that are not co-visible in the same frame. We evaluated seven state-of-the-art VLMs and found that models lag behind human performance by 28%, with steeper declines in accuracy (60% to 30 %) as the temporal gap widens. Our analysis further reveals that providing trajectories or bird's-eye-view projections to VLMs results in only marginal improvements, whereas providing oracle 3D coordinates leads to a substantial 20% performance increase. This highlights a core bottleneck of multi-frame VLMs in constructing and maintaining 3D scene representations over time from visual signals. Disjoint-3DQA therefore sets a clear, measurable challenge for long-horizon spatial reasoning and aims to catalyze future research at the intersection of vision, language, and embodied AI.