WISER：面向免训练零样本组合图像检索的更广搜索、更深思考与自适应融合
WISER: Wider Search, Deeper Thinking, and Adaptive Fusion for Training-Free Zero-Shot Composed Image Retrieval
提出免训练的多模态检索框架，通过并行生成编辑文本和编辑图像扩大候选池，并用验证器评估置信度触发自反思精炼，可复用于其他需要融合多模态查询的检索任务。
原文依据 we propose WISER, a training-free framework that unifies T2I and I2I via a "retrieve-verify-refine" pipeline
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论零样本组合图像检索（ZS-CIR）旨在给定多模态查询（包含参考图像和修改文本）的情况下检索目标图像，且无需在标注三元组上训练。现有方法通常将多模态查询转换为单一模态——要么作为编辑后的描述用于文本到图像检索（T2I），要么作为编辑后的图像用于图像到图像检索（I2I）。然而，T2I 常常丢失细粒度视觉细节，而 I2I 难以处理复杂的语义修改。为了在不同查询意图下有效利用它们的互补优势，我们提出了 WISER，一个免训练框架，通过“检索-验证-精炼”流程统一 T2I 和 I2I，显式建模意图感知和不确定性感知。具体而言，WISER 首先执行更广搜索，生成编辑后的描述和图像进行并行检索以扩大候选池。然后，它通过验证器进行自适应融合，评估检索置信度，对不确定的检索触发精炼，并对可靠的检索动态融合双路径。对于不确定的检索，WISER 通过结构化自反思生成精炼建议，以引导下一轮检索走向更深思考。大量实验表明，WISER 在多个基准上具有优越性能，相较于现有免训练方法，在 CIRCO（mAP@5）上取得 45% 的相对提升，在 CIRR（Recall@1）上取得 57% 的相对提升。代码已发布于 https://github.com/Physicsmile/WISER。
英文原文摘要 1522 字符
Zero-Shot Composed Image Retrieval (ZS-CIR) aims to retrieve target images given a multimodal query (comprising a reference image and a modification text), without training on annotated triplets. Existing methods typically convert the multimodal query into a single modality--either as an edited caption for Text-to-Image retrieval (T2I) or as an edited image for Image-to-Image retrieval (I2I). However, T2I often loses fine-grained visual details, while I2I struggles with complex semantic modifications. To effectively leverage their complementary strengths under diverse query intents, we propose WISER, a training-free framework that unifies T2I and I2I via a "retrieve-verify-refine" pipeline, explicitly modeling intent awareness and uncertainty awareness. Specifically, WISER first performs Wider Search by generating both edited captions and images for parallel retrieval to broaden the candidate pool. Then, it conducts Adaptive Fusion with a verifier to assess retrieval confidence, triggering refinement for uncertain retrievals, and dynamically fusing the dual-path for reliable ones. For uncertain retrievals, WISER generates refinement suggestions through structured self-reflection to guide the next retrieval round toward Deeper Thinking. Extensive experiments show the superior performance of WISER across multiple benchmarks, achieving relative improvements of 45% on CIRCO (mAP@5) and 57% on CIRR (Recall@1) over existing training-free methods. Code is released at https://github.com/Physicsmile/WISER.