Towards Comprehensive Scene Understanding: Integrating First and Third-Person Views for LVLMs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lee, Insu, Park, Wooje, Jang, Jaeyun, Noh, Minyoung, Shim, Kyuhong, Shim, Byonghyo
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911228847194112
author Lee, Insu
Park, Wooje
Jang, Jaeyun
Noh, Minyoung
Shim, Kyuhong
Shim, Byonghyo
author_facet Lee, Insu
Park, Wooje
Jang, Jaeyun
Noh, Minyoung
Shim, Kyuhong
Shim, Byonghyo
contents Large vision-language models (LVLMs) are increasingly deployed in interactive applications such as virtual and augmented reality, where a first-person (egocentric) view captured by head-mounted cameras serves as key input. While this view offers fine-grained cues about user attention and hand-object interactions, its narrow field of view and lack of global context often lead to failures on spatially or contextually demanding queries. To address this, we introduce a framework that augments egocentric inputs with third-person (exocentric) views, providing complementary information such as global scene layout and object visibility to LVLMs. We present E3VQA, the first benchmark for multi-view question answering with 4K high-quality question-answer pairs grounded in synchronized ego-exo image pairs. Additionally, we propose M3CoT, a training-free prompting technique that constructs a unified scene representation by integrating scene graphs from three complementary perspectives. M3CoT enables LVLMs to reason more effectively across views, yielding consistent performance gains (4.84% for GPT-4o and 5.94% for Gemini 2.0 Flash) over a recent CoT baseline. Our extensive evaluation reveals key strengths and limitations of LVLMs in multi-view reasoning and highlights the value of leveraging both egocentric and exocentric inputs. The dataset and source code are available at https://github.com/Leeinsu1/Towards-Comprehensive-Scene-Understanding.
format Preprint
id arxiv_https___arxiv_org_abs_2505_21955
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards Comprehensive Scene Understanding: Integrating First and Third-Person Views for LVLMs
Lee, Insu
Park, Wooje
Jang, Jaeyun
Noh, Minyoung
Shim, Kyuhong
Shim, Byonghyo
Computer Vision and Pattern Recognition
Artificial Intelligence
Large vision-language models (LVLMs) are increasingly deployed in interactive applications such as virtual and augmented reality, where a first-person (egocentric) view captured by head-mounted cameras serves as key input. While this view offers fine-grained cues about user attention and hand-object interactions, its narrow field of view and lack of global context often lead to failures on spatially or contextually demanding queries. To address this, we introduce a framework that augments egocentric inputs with third-person (exocentric) views, providing complementary information such as global scene layout and object visibility to LVLMs. We present E3VQA, the first benchmark for multi-view question answering with 4K high-quality question-answer pairs grounded in synchronized ego-exo image pairs. Additionally, we propose M3CoT, a training-free prompting technique that constructs a unified scene representation by integrating scene graphs from three complementary perspectives. M3CoT enables LVLMs to reason more effectively across views, yielding consistent performance gains (4.84% for GPT-4o and 5.94% for Gemini 2.0 Flash) over a recent CoT baseline. Our extensive evaluation reveals key strengths and limitations of LVLMs in multi-view reasoning and highlights the value of leveraging both egocentric and exocentric inputs. The dataset and source code are available at https://github.com/Leeinsu1/Towards-Comprehensive-Scene-Understanding.
title Towards Comprehensive Scene Understanding: Integrating First and Third-Person Views for LVLMs
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2505.21955