Beyond Semantic Search: Towards Referential Anchoring in Composed Image Retrieval

TL;DR

This paper introduces Object-Anchored Composed Image Retrieval (OACIR), a new fine-grained retrieval task emphasizing instance fidelity, supported by a large-scale benchmark and a novel attention-based framework called AdaFocal.

Contribution

It proposes OACIR as a new task, creates the first large-scale benchmark OACIRR, and develops AdaFocal, an attention framework that enhances instance-level retrieval accuracy.

Findings

AdaFocal outperforms existing models in maintaining instance fidelity.

OACIRR benchmark contains over 160K quadruples with hard-negative distractors.

AdaFocal effectively balances attention between the anchored object and context.

Abstract

Composed Image Retrieval (CIR) has demonstrated significant potential by enabling flexible multimodal queries that combine a reference image and modification text. However, CIR inherently prioritizes semantic matching, struggling to reliably retrieve a user-specified instance across contexts. In practice, emphasizing concrete instance fidelity over broad semantics is often more consequential. In this work, we propose Object-Anchored Composed Image Retrieval (OACIR), a novel fine-grained retrieval task that mandates strict instance-level consistency. To advance research on this task, we construct OACIRR (OACIR on Real-world images), the first large-scale, multi-domain benchmark comprising over 160K quadruples and four challenging candidate galleries enriched with hard-negative instance distractors. Each quadruple augments the compositional query with a bounding box that visually anchors the object in the reference image, providing a precise and flexible way to ensure instance preservation. To address the OACIR task, we propose AdaFocal, a framework featuring a Context-Aware Attention Modulator that adaptively intensifies attention within the specified instance region, dynamically balancing focus between the anchored instance and the broader compositional context. Extensive experiments demonstrate that AdaFocal substantially outperforms existing compositional retrieval models, particularly in maintaining instance-level fidelity, thereby establishing a robust baseline for this challenging task while opening new directions for more flexible, instance-aware retrieval systems.