LLM Meets Scene Graph: Can Large Language Models Understand and Generate Scene Graphs? A Benchmark and Empirical Study

TL;DR

This paper introduces TSG Bench, a benchmark for evaluating large language models' ability to understand and generate scene graphs from text, revealing strengths in understanding but challenges in complex scene graph generation.

Contribution

The paper presents TSG Bench, a new benchmark for assessing LLMs' scene graph understanding and generation, along with an empirical evaluation of 11 models highlighting current limitations.

Findings

Models excel at understanding scene graphs.

Models struggle with generating scene graphs from complex narratives.

Generation bottleneck identified in decomposing scenes.

Abstract

The remarkable reasoning and generalization capabilities of Large Language Models (LLMs) have paved the way for their expanding applications in embodied AI, robotics, and other real-world tasks. To effectively support these applications, grounding in spatial and temporal understanding in multimodal environments is essential. To this end, recent works have leveraged scene graphs, a structured representation that encodes entities, attributes, and their relationships in a scene. However, a comprehensive evaluation of LLMs' ability to utilize scene graphs remains limited. In this work, we introduce Text-Scene Graph (TSG) Bench, a benchmark designed to systematically assess LLMs' ability to (1) understand scene graphs and (2) generate them from textual narratives. With TSG Bench we evaluate 11 LLMs and reveal that, while models perform well on scene graph understanding, they struggle with scene graph generation, particularly for complex narratives. Our analysis indicates that these models fail to effectively decompose discrete scenes from a complex narrative, leading to a bottleneck when generating scene graphs. These findings underscore the need for improved methodologies in scene graph generation and provide valuable insights for future research. The demonstration of our benchmark is available at https://tsg-bench.netlify.app. Additionally, our code and evaluation data are publicly available at https://github.com/docworlds/tsg-bench.