arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.20972cs.CV
收藏

多模态大语言模型在图像标签任务中是好的标注者吗?

Are Multimodal Large Language Models Good Annotators for Image Tagging?

RIKEN Center for Advanced Intelligence Project, Japan(日本先进人工智能研究中心);The University of Tokyo, Japan(东京大学);Southeast University, China(东南大学);China University of Mining(中国矿业大学)

Ming-Kun Xie, Jia-Hao Xiao, Zhiqiang Kou, Zhongnian Li, Gang Niu, Masashi Sugiyama

AI总结:

本文提出TagLLM框架,通过候选生成和标签歧义消除方法,有效缩小多模态大语言模型生成标注与人工标注之间的差距,提升图像标签任务的效率和性能。

摘要:

图像标签是一项基本的视觉任务,传统上依赖于人工标注的数据集来训练多标签分类器,这会带来显著的劳动和成本。虽然多模态大语言模型(MLLMs)提供了自动化标注的潜力,但它们取代人工标注的能力仍缺乏深入研究。本文旨在分析MLLM生成标注与人工标注之间的差距,并提出一个有效的解决方案,使基于MLLM的标注能够取代手动标注。我们对MLLM标注的分析表明,在保守估计下,MLLMs可以将标注成本降低到人工成本的千分之一,主要体现在GPU使用上,这与人工努力相比几乎可以忽略不计。它们的标注质量达到约人类性能的50%至80%,而在下游训练任务中达到超过90%的性能。受这些发现的启发,我们提出了TagLLM,一个新颖的图像标签框架,旨在缩小MLLM生成标注与人工标注之间的差距。TagLLM包含两个组件:候选生成,它利用结构化的组内提示法,高效地生成一个紧凑的候选集,尽可能覆盖更多的真实标签,同时减少后续标注的工作量;以及标签歧义消除,它通过交互式校准提示中的语义概念,有效细化候选标签。广泛的实验表明,TagLLM显著缩小了MLLM生成标注与人工标注之间的差距,尤其是在下游训练性能方面,它缩小了约60%至80%的差距。

英文摘要:

Image tagging, a fundamental vision task, traditionally relies on human-annotated datasets to train multi-label classifiers, which incurs significant labor and costs. While Multimodal Large Language Models (MLLMs) offer promising potential to automate annotation, their capability to replace human annotators remains underexplored. This paper aims to analyze the gap between MLLM-generated and human annotations and to propose an effective solution that enables MLLM-based annotation to replace manual labeling. Our analysis of MLLM annotations reveals that, under a conservative estimate, MLLMs can reduce annotation cost to as low as one-thousandth of the human cost, mainly accounting for GPU usage, which is nearly negligible compared to manual efforts. Their annotation quality reaches about 50\% to 80\% of human performance, while achieving over 90\% performance on downstream training tasks.Motivated by these findings, we propose TagLLM, a novel framework for image tagging, which aims to narrow the gap between MLLM-generated and human annotations. TagLLM comprises two components: Candidates generation, which employs structured group-wise prompting to efficiently produce a compact candidate set that covers as many true labels as possible while reducing subsequent annotation workload; and label disambiguation, which interactively calibrates the semantic concept of categories in the prompts and effectively refines the candidate labels. Extensive experiments show that TagLLM substantially narrows the gap between MLLM-generated and human annotations, especially in downstream training performance, where it closes about 60\% to 80\% of the difference.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑