Skip to content
AI News HubLIVE
Original source2 min read

Can Vision-Language Models Assess Proxemic Risk from Egocentric Robot Images?

Summary

A study accepted at the EMR 2026 workshop at ECCV 2026 evaluates three open-source vision-language models—InternVL, Qwen-VL, and SmolVLM—on classifying egocentric robot images into four proxemic danger levels. Without fine-tuning, all models perform near a stratified random baseline; two rounds of QLoRA fine-tuning yield only modest overall gains, though Qwen-VL with an advanced prompt achieves substantially higher recall for high-danger cases. The analysis also finds that correct danger classification does not imply better spatial grounding of people.

SourcearXiv Computer VisionAuthor: Vladyslava Rudas, Dmytro Kuzmenko
Can Vision-Language Models Assess Proxemic Risk from Egocentric Robot Images?
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 12 Aug 2026]

Title:Can Vision-Language Models Assess Proxemic Risk from Egocentric Robot Images?

View a PDF of the paper titled Can Vision-Language Models Assess Proxemic Risk from Egocentric Robot Images?, by Vladyslava Rudas and Dmytro Kuzmenko

View PDF HTML (experimental)

Abstract:Assessing proxemic danger from a robot's egocentric perspective is critical for safe embodied navigation in human environments and requires both visual and contextual reasoning. We evaluate three opensource vision-language models (VLMs) (\textit{InternVL}, \textit{Qwen-VL}, and \textit{SmolVLM}) on the classification of egocentric robot images into four danger levels, comparing three prompting strategies and two rounds of QLoRA fine-tuning against a stratified random baseline. Without fine-tuning, all models perform near the baseline, while fine-tuning yields only modest overall improvements. However, \textit{Qwen-VL} with an advanced prompt achieves substantially higher recall for high-danger cases than the other models. An analysis of person localization further shows that correct danger classification does not correspond to better spatial grounding, indicating that a model may produce a useful safety label without attending to the relevant region of the scene. These results show that current VLMs remain limited in fine-grained proxemic reasoning and spatial grounding, although targeted prompting and fine-tuning can improve high-danger detection in selected models.

Comments: Accepted at the EMR 2026 workshop at ECCV 2026 (non-archival)

Subjects:

Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)

Cite as: arXiv:2608.12515 [cs.CV]

(or arXiv:2608.12515v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2608.12515

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Dmytro Kuzmenko [view email] [v1] Wed, 12 Aug 2026 18:47:43 UTC (33 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Can Vision-Language Models Assess Proxemic Risk from Egocentric Robot Images?, by Vladyslava Rudas and Dmytro Kuzmenko

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.CV

new | recent | 2026-08

Change to browse by:

cs cs.RO

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersIntermediate

Key points

  • The study compares InternVL, Qwen-VL, and SmolVLM on four-level proxemic danger classification from robot egocentric images.
  • Without fine-tuning, all models perform near a stratified random baseline, and two rounds of QLoRA fine-tuning provide only modest overall improvements.
  • Qwen-VL with an advanced prompt achieves substantially higher recall on high-danger cases than the other models.
  • Correct danger classification does not correspond to better person localization, suggesting models may produce useful safety labels without attending to the relevant scene regions.

Highlights and analysis are generated automatically and may contain errors. Check the original source.