Open World Scene Graph Generation using Vision Language Models

TL;DR

This paper presents an open-world scene graph generation method that leverages pretrained vision-language models in a zero-shot, training-free manner, enabling relational understanding of images with unseen objects and relations.

Contribution

It introduces a novel, training-free framework that uses multimodal prompting and embedding alignment to perform scene graph generation without dataset-specific fine-tuning.

Findings

Effective zero-shot performance on multiple datasets

Capable of handling unseen objects and relations

Outperforms traditional supervised methods in open-world scenarios

Abstract

Scene-Graph Generation (SGG) seeks to recognize objects in an image and distill their salient pairwise relationships. Most methods depend on dataset-specific supervision to learn the variety of interactions, restricting their usefulness in open-world settings, involving novel objects and/or relations. Even methods that leverage large Vision Language Models (VLMs) typically require benchmark-specific fine-tuning. We introduce Open-World SGG, a training-free, efficient, model-agnostic framework that taps directly into the pretrained knowledge of VLMs to produce scene graphs with zero additional learning. Casting SGG as a zero-shot structured-reasoning problem, our method combines multimodal prompting, embedding alignment, and a lightweight pair-refinement strategy, enabling inference over unseen object vocabularies and relation sets. To assess this setting, we formalize an Open-World evaluation protocol that measures performance when no SGG-specific data have been observed either in terms of objects and relations. Experiments on Visual Genome, Open Images V6, and the Panoptic Scene Graph (PSG) dataset demonstrate the capacity of pretrained VLMs to perform relational understanding without task-level training.