Knowledge-Enhanced Dual-stream Zero-shot Composed Image Retrieval

TL;DR

This paper introduces KEDs, a knowledge-enhanced dual-stream framework for zero-shot composed image retrieval that models detailed attributes and aligns visual and textual semantics, outperforming previous methods.

Contribution

The paper proposes a novel dual-stream framework that incorporates a database for attribute modeling and aligns pseudo-word tokens with textual concepts in a zero-shot setting.

Findings

KEDs outperforms previous zero-shot CIR methods on multiple benchmarks.

The framework effectively models detailed attributes like color and layout.

Explicit alignment of visual tokens with text improves retrieval accuracy.

Abstract

We study the zero-shot Composed Image Retrieval (ZS-CIR) task, which is to retrieve the target image given a reference image and a description without training on the triplet datasets. Previous works generate pseudo-word tokens by projecting the reference image features to the text embedding space. However, they focus on the global visual representation, ignoring the representation of detailed attributes, e.g., color, object number and layout. To address this challenge, we propose a Knowledge-Enhanced Dual-stream zero-shot composed image retrieval framework (KEDs). KEDs implicitly models the attributes of the reference images by incorporating a database. The database enriches the pseudo-word tokens by providing relevant images and captions, emphasizing shared attribute information in various aspects. In this way, KEDs recognizes the reference image from diverse perspectives. Moreover, KEDs adopts an extra stream that aligns pseudo-word tokens with textual concepts, leveraging pseudo-triplets mined from image-text pairs. The pseudo-word tokens generated in this stream are explicitly aligned with fine-grained semantics in the text embedding space. Extensive experiments on widely used benchmarks, i.e. ImageNet-R, COCO object, Fashion-IQ and CIRR, show that KEDs outperforms previous zero-shot composed image retrieval methods.