CSMCIR: CoT-Enhanced Symmetric Alignment with Memory Bank for Composed Image Retrieval

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Qian, Zhipeng, Liang, Zihan, Ma, Yufei, Chen, Ben, Dai, Huangyu, Ma, Yiwei, Ji, Jiayi, Lei, Chenyi, Li, Han, Sun, Xiaoshuai
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915990289252352
author Qian, Zhipeng
Liang, Zihan
Ma, Yufei
Chen, Ben
Dai, Huangyu
Ma, Yiwei
Ji, Jiayi
Lei, Chenyi
Li, Han
Sun, Xiaoshuai
author_facet Qian, Zhipeng
Liang, Zihan
Ma, Yufei
Chen, Ben
Dai, Huangyu
Ma, Yiwei
Ji, Jiayi
Lei, Chenyi
Li, Han
Sun, Xiaoshuai
contents Composed Image Retrieval (CIR) enables users to search for target images using both a reference image and manipulation text, offering substantial advantages over single-modality retrieval systems. However, existing CIR methods suffer from representation space fragmentation: queries and targets comprise heterogeneous modalities and are processed by distinct encoders, forcing models to bridge misaligned representation spaces only through post-hoc alignment, which fundamentally limits retrieval performance. This architectural asymmetry manifests as three distinct, well-separated clusters in the feature space, directly demonstrating how heterogeneous modalities create fundamentally misaligned representation spaces from initialization. In this work, we propose CSMCIR, a unified representation framework that achieves efficient query-target alignment through three synergistic components. First, we introduce a Multi-level Chain-of-Thought (MCoT) prompting strategy that guides Multimodal Large Language Models to generate discriminative, semantically compatible captions for target images, establishing modal symmetry. Building upon this, we design a symmetric dual-tower architecture where both query and target sides utilize the identical shared-parameter Q-Former for cross-modal encoding, ensuring consistent feature representations and further reducing the alignment gap. Finally, this architectural symmetry enables an entropy-based, temporally dynamic Memory Bank strategy that provides high-quality negative samples while maintaining consistency with the evolving model state. Extensive experiments on four benchmark datasets demonstrate that our CSMCIR achieves state-of-the-art performance with superior training efficiency. Comprehensive ablation studies further validate the effectiveness of each proposed component.
format Preprint
id arxiv_https___arxiv_org_abs_2601_03728
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle CSMCIR: CoT-Enhanced Symmetric Alignment with Memory Bank for Composed Image Retrieval
Qian, Zhipeng
Liang, Zihan
Ma, Yufei
Chen, Ben
Dai, Huangyu
Ma, Yiwei
Ji, Jiayi
Lei, Chenyi
Li, Han
Sun, Xiaoshuai
Computer Vision and Pattern Recognition
Artificial Intelligence
Composed Image Retrieval (CIR) enables users to search for target images using both a reference image and manipulation text, offering substantial advantages over single-modality retrieval systems. However, existing CIR methods suffer from representation space fragmentation: queries and targets comprise heterogeneous modalities and are processed by distinct encoders, forcing models to bridge misaligned representation spaces only through post-hoc alignment, which fundamentally limits retrieval performance. This architectural asymmetry manifests as three distinct, well-separated clusters in the feature space, directly demonstrating how heterogeneous modalities create fundamentally misaligned representation spaces from initialization. In this work, we propose CSMCIR, a unified representation framework that achieves efficient query-target alignment through three synergistic components. First, we introduce a Multi-level Chain-of-Thought (MCoT) prompting strategy that guides Multimodal Large Language Models to generate discriminative, semantically compatible captions for target images, establishing modal symmetry. Building upon this, we design a symmetric dual-tower architecture where both query and target sides utilize the identical shared-parameter Q-Former for cross-modal encoding, ensuring consistent feature representations and further reducing the alignment gap. Finally, this architectural symmetry enables an entropy-based, temporally dynamic Memory Bank strategy that provides high-quality negative samples while maintaining consistency with the evolving model state. Extensive experiments on four benchmark datasets demonstrate that our CSMCIR achieves state-of-the-art performance with superior training efficiency. Comprehensive ablation studies further validate the effectiveness of each proposed component.
title CSMCIR: CoT-Enhanced Symmetric Alignment with Memory Bank for Composed Image Retrieval
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2601.03728