arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2505.23769cs.CV

TextRegion:来自冻结图像-文本模型的文本对齐区域令牌

TextRegion: Text-Aligned Region Tokens from Frozen Image-Text Models

Siebel School of Computing and Data Science(塞比尔计算与数据科学学院);University of Illinois at Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校)

Yao Xiao, Qiqian Fu, Heyi Tao, Yuqun Wu, Zhen Zhu, Derek Hoiem

AI总结:

本文提出无需训练的TextRegion框架,结合图像-文本模型与SAM2分割模型生成文本对齐区域令牌,在开放世界语义分割、指代表达理解等任务上取得优越性能。

摘要:

图像-文本模型在图像级任务上表现出色,但在细粒度视觉理解方面存在不足。虽然这些模型提供了强大的视觉-语言对齐能力,但像SAM2这样的分割模型能够为对象提供精确的空间边界。为此,我们提出了TextRegion,一个简单、有效且无需训练的框架,它结合了图像-文本模型和SAM2的优势,生成强大的文本对齐区域令牌。这些令牌在保留开放词汇能力的同时,实现了细粒度的视觉理解。它们可以直接应用于各种下游任务,包括开放世界语义分割、指代表达理解和定位。我们进行了广泛的评估,与最先进的无需训练方法相比,始终取得了优越或有竞争力的性能。此外,我们的框架与许多图像-文本模型兼容,随着更强模型的出现,它具有高度的实用性和易于扩展性。代码可在以下地址获取:https://github.com/avaxiao/TextRegion。

英文摘要:

Image-text models excel at image-level tasks but struggle with detailed visual understanding. While these models provide strong visual-language alignment, segmentation models like SAM2 offer precise spatial boundaries for objects. To this end, we propose TextRegion, a simple, effective, and training-free framework that combines the strengths of image-text models and SAM2 to generate powerful text-aligned region tokens. These tokens enable detailed visual understanding while preserving open-vocabulary capabilities. They can be directly applied to various downstream tasks, including open-world semantic segmentation, referring expression comprehension, and grounding. We conduct extensive evaluations and consistently achieve superior or competitive performance compared to state-of-the-art training-free methods. Additionally, our framework is compatible with many image-text models, making it highly practical and easily extensible as stronger models emerge. Code is available at: https://github.com/avaxiao/TextRegion.

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。

↑