HyperGLM: HyperGraph for Video Scene Graph Generation and Anticipation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Nguyen, Trong-Thuan, Nguyen, Pha, Cothren, Jackson, Yilmaz, Alper, Luu, Khoa
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908291247898624
author Nguyen, Trong-Thuan
Nguyen, Pha
Cothren, Jackson
Yilmaz, Alper
Luu, Khoa
author_facet Nguyen, Trong-Thuan
Nguyen, Pha
Cothren, Jackson
Yilmaz, Alper
Luu, Khoa
contents Multimodal LLMs have advanced vision-language tasks but still struggle with understanding video scenes. To bridge this gap, Video Scene Graph Generation (VidSGG) has emerged to capture multi-object relationships across video frames. However, prior methods rely on pairwise connections, limiting their ability to handle complex multi-object interactions and reasoning. To this end, we propose Multimodal LLMs on a Scene HyperGraph (HyperGLM), promoting reasoning about multi-way interactions and higher-order relationships. Our approach uniquely integrates entity scene graphs, which capture spatial relationships between objects, with a procedural graph that models their causal transitions, forming a unified HyperGraph. Significantly, HyperGLM enables reasoning by injecting this unified HyperGraph into LLMs. Additionally, we introduce a new Video Scene Graph Reasoning (VSGR) dataset featuring 1.9M frames from third-person, egocentric, and drone views and supports five tasks: Scene Graph Generation, Scene Graph Anticipation, Video Question Answering, Video Captioning, and Relation Reasoning. Empirically, HyperGLM consistently outperforms state-of-the-art methods across five tasks, effectively modeling and reasoning complex relationships in diverse video scenes.
format Preprint
id arxiv_https___arxiv_org_abs_2411_18042
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle HyperGLM: HyperGraph for Video Scene Graph Generation and Anticipation
Nguyen, Trong-Thuan
Nguyen, Pha
Cothren, Jackson
Yilmaz, Alper
Luu, Khoa
Computer Vision and Pattern Recognition
Multimodal LLMs have advanced vision-language tasks but still struggle with understanding video scenes. To bridge this gap, Video Scene Graph Generation (VidSGG) has emerged to capture multi-object relationships across video frames. However, prior methods rely on pairwise connections, limiting their ability to handle complex multi-object interactions and reasoning. To this end, we propose Multimodal LLMs on a Scene HyperGraph (HyperGLM), promoting reasoning about multi-way interactions and higher-order relationships. Our approach uniquely integrates entity scene graphs, which capture spatial relationships between objects, with a procedural graph that models their causal transitions, forming a unified HyperGraph. Significantly, HyperGLM enables reasoning by injecting this unified HyperGraph into LLMs. Additionally, we introduce a new Video Scene Graph Reasoning (VSGR) dataset featuring 1.9M frames from third-person, egocentric, and drone views and supports five tasks: Scene Graph Generation, Scene Graph Anticipation, Video Question Answering, Video Captioning, and Relation Reasoning. Empirically, HyperGLM consistently outperforms state-of-the-art methods across five tasks, effectively modeling and reasoning complex relationships in diverse video scenes.
title HyperGLM: HyperGraph for Video Scene Graph Generation and Anticipation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.18042