arXiv cs.AI· Yuqian Zheng, Hyoju Ahn, Hyeongjin Ahn, Eunil Park·· 12 小时前AI 评分18
Less-is-More CIR：面向组合图像检索的简洁检索与 Critic 重排序框架
Less is More: Concise Retrieval and Critic-based Re-ranking for Composed Image Retrieval
AI 导读
Less-is-More CIR 是一个无需训练的两阶段框架，用于零样本组合图像检索（ZS-CIR）：第一阶段只生成保留最具区分度修改语义的简洁目标导向描述，减少检索中的语义噪声；第二阶段用 critic 重排序模块从修改一致性和参考图保留两个角度评估候选结果。该工作已被 CIKM 2026 接收，代码已公开。
正文 · 原文
Abstract:Zero-shot composed image retrieval (ZS-CIR) aims to retrieve a target image based on a reference image and modification text. Recent approaches commonly rely on large vision-language models to generate target image captions for retrieval. However, these captions may include redundant semantic details, introducing noise when retrieving the target image and leading to suboptimal retrieval results. To address this issue, we propose Less-is-More CIR, a two-stage training-free framework that combines Concise Retrieval and Critic Re-ranking. In the first stage, concise target-oriented captions are generated that preserve only the most discriminative modification semantics, thereby reducing semantic noise in the retrieval. In the second stage, a critic re-rank module further evaluates the retrieved candidates from the perspectives of modification consistency and reference preservation. Extensive experiments demonstrate the effectiveness and robustness of the proposed framework for the ZS-CIR. Our code is available at: this https URL.
| Comments: | Accepted at CIKM 2026 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.27538 [cs.CV] |
| (or arXiv:2610.27538v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.27538 arXiv-issued DOI via DataCite (pending registration) |
|
| Related DOI: | https://doi.org/10.1145/3799682.3840662
DOI(s) linking to related resources |
Submission history
From: Yuqian Zheng [view email]
[v1]
Tue, 20 Oct 2026 02:53:02 UTC (9,983 KB)
来源：arXiv cs.AI · arxiv.org