CVMMOct 20, 2024

Scene Graph Generation with Role-Playing Large Language Models

arXiv:2410.15364v128 citationsh-index: 7NIPS
Originality Incremental advance
AI Analysis

This work addresses the challenge of generating scene graphs with open vocabularies for computer vision applications, representing an incremental advance by refining text classifier adaptation and relation recognition.

The paper tackles the problem of open-vocabulary scene graph generation by addressing the limitations of scene-agnostic text classifiers in existing methods, proposing SDSGG, a framework that adaptively adjusts text classifiers based on visual content and uses role-playing LLMs to generate diverse scene descriptions, resulting in clear performance improvements over top-leading methods on benchmarks.

Current approaches for open-vocabulary scene graph generation (OVSGG) use vision-language models such as CLIP and follow a standard zero-shot pipeline -- computing similarity between the query image and the text embeddings for each category (i.e., text classifiers). In this work, we argue that the text classifiers adopted by existing OVSGG methods, i.e., category-/part-level prompts, are scene-agnostic as they remain unchanged across contexts. Using such fixed text classifiers not only struggles to model visual relations with high variance, but also falls short in adapting to distinct contexts. To plug these intrinsic shortcomings, we devise SDSGG, a scene-specific description based OVSGG framework where the weights of text classifiers are adaptively adjusted according to the visual content. In particular, to generate comprehensive and diverse descriptions oriented to the scene, an LLM is asked to play different roles (e.g., biologist and engineer) to analyze and discuss the descriptive features of a given scene from different views. Unlike previous efforts simply treating the generated descriptions as mutually equivalent text classifiers, SDSGG is equipped with an advanced renormalization mechanism to adjust the influence of each text classifier based on its relevance to the presented scene (this is what the term "specific" means). Furthermore, to capture the complicated interplay between subjects and objects, we propose a new lightweight module called mutual visual adapter. It refines CLIP's ability to recognize relations by learning an interaction-aware semantic space. Extensive experiments on prevalent benchmarks show that SDSGG outperforms top-leading methods by a clear margin.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes