COMBINER: Composed Image Retrieval Guided by Attribute-based Neighbor Relations

2026-06-03Computer Vision and Pattern Recognition

Computer Vision and Pattern Recognition
AI summary

The authors address a problem in composed image retrieval, where finding the right image is hard when pictures look similar but have different attributes. They propose a method called COMBINER that breaks down and combines image and text features using attribute prototypes to better understand these differences. Their approach also models relationships between images based on attributes to improve matching accuracy. Tests on standard datasets show their method works well compared to previous ones.

Composed Image RetrievalMultimodal FeaturesAttribute PrototypesSemantic DisentanglementCross-modal RepresentationSimilarity ModelingNeighbor RelationsMultimodal Feature CompositionBenchmark Datasets
Authors
Zixu Li, Yupeng Hu, Zhiwei Chen, Haokun Wen, Xuemeng Song, Liqiang Nie
Abstract
Composed Image Retrieval (CIR) represents a challenging retrieval task that targets locating specific images through multimodal inputs. Despite recent progress in CIR techniques, prior approaches often overlook cases where images appear visually alike yet differ in attributes, potentially undermining both multimodal feature fusion and similarity modeling. To mitigate this limitation, we design a unified representation of cross-modal features based on attribute prototypes. Nevertheless, the task is far from straightforward, owing to three core issues: (1) entanglement in attribute-level semantics, (2) inconsistency across modalities, and (3) supervised signal missing. To tackle the above obstacles, we introduce a COMposed image retrieval network guided By attrIbute-based NEighbor Relations (COMBINER). Specifically, we first design an Adaptive Semantic Disentanglement module, which is capable of disentangling attribute features based on multimodal primitive features. Secondly, we propose a Unified Prototype-based Composition module, which can construct cross-modal unified prototypes (CUP) and facilitate multimodal feature composition. Finally, we introduce a Dual Relations Modeling module, which can mine pairwise and neighbor relations based on attribute similarity. Compared to traditional neighbor relations modeling CIR methods, COMBINER represents the first study addressing the phenomenon of visually similar but attribute-unrelated samples. It achieves a more accurate understanding of the semantic relations among samples by employing an attribute prototype-based similarity metric. Comprehensive experiments conducted on three benchmark datasets confirm the effectiveness of our proposed COMBINER. The implementation of our method will be accessed at https://github.com/Lee-zixu/COMBINER