CaptionFuse: Training-Free Composed Product Retrieval Via VLM Intent Captioning and Weighted Embedding Fusion

Zhao Gao, Mina Metias, Mohamed Noordeen Alaudeen, Bassel Saleh
10/19/2026
cs.CV

No confident code implementation found for this paper.

Abstract

Composed Image Retrieval (CIR) retrieves products matching a reference image modified by natural language---a core capability for eCommerce search. Existing training-free methods fuse CLIP embeddings via arithmetic, but we identify a fundamental limitation: user text modifications are attribute-sparse---a user who types ``make it red with shorter sleeves'' specifies what to change but not what to keep (the floral pattern, the V-neck, the fit-and-flare silhouette). We introduce language-space fusion: a Vision-Language Model (VLM) jointly interprets the reference image and modification to generate a visually grounded product-title caption---e.g., ``Red floral fit-and-flare dress with short sleeves and V-neck''---that fills in the missing attributes from the image. Our method, CaptionFuse, requires no task-specific training and is model-agnostic. A controlled experiment confirms that attribute content--- not linguistic style---drives the gains: imperative and descriptive captions with identical attributes achieve equal R@10, while both outperform raw modifications by +6.5 points.

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

Exploring Multi-Modal Large Language Models and Two-Stage Fine-Tuning for Fashion Image Retrieval

86% match2026

Scaling Prompt Instructed Zero Shot Composed Image Retrieval with Image-Only Data

84% match2025

ProductConsistency: Improving Product Identity Preservation in Instruction-Based Image Editing via SFT and RL

83% match2026

Edit Transfer: Learning Image Editing via Vision In-Context Relations

83% match1 repo902025

CoTMR: Chain-of-Thought Multi-Scale Reasoning for Training-Free Zero-Shot Composed Image Retrieval

82% match2025

Evcap: Retrieval-Augmented Image Captioning with External Visual-Name Memory for Open-World Comprehension

82% match1 repo632026

Cite this paper

@article{gao2026captionfuse,
  title  = {CaptionFuse: Training-Free Composed Product Retrieval Via VLM Intent Captioning and Weighted Embedding Fusion},
  author = {Zhao Gao and Mina Metias and Mohamed Noordeen Alaudeen and Bassel Saleh},
  year   = {2026},
  eprint = {2610.26661},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2610.26661v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2610.26661.svg)](https://codeofpaper.com/paper/2610.26661)

Discussion