Search papers, labs, and topics across Lattice.
This paper introduces Object-Anchored Composed Image Retrieval (OACIR), a new task that emphasizes instance-level consistency in composed image retrieval by incorporating a bounding box to visually anchor the object in the reference image. To facilitate research, they created OACIRR, a large-scale, multi-domain benchmark with over 160K quadruples and challenging candidate galleries. They also propose AdaFocal, a framework with a Context-Aware Attention Modulator, which significantly outperforms existing models in maintaining instance-level fidelity, establishing a strong baseline for OACIR.
Forget semantic similarity – now you can pinpoint *exactly* which object to retrieve in a composed image search with bounding box anchors, thanks to a new benchmark and model that blows away existing methods.
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.
