Mask-aware Text-to-Image Retrieval: Referring Expression Segmentation Meets Cross-modal Retrieval

TL;DR

This paper introduces Mask-aware TIR (MaTIR), a unified framework combining efficient text-to-image retrieval with precise object segmentation, leveraging large language models and region-level embeddings for improved accuracy and interpretability.

Contribution

The paper proposes a novel two-stage framework that unifies TIR and RES, utilizing SAM 2, Alpha-CLIP, and MLLM for scalable, accurate retrieval and segmentation.

Findings

Significant improvements in retrieval accuracy on COCO and D$^3$ datasets.

Enhanced segmentation quality with efficient object localization.

Effective integration of large language models for reranking and grounding.

Abstract

Text-to-image retrieval (TIR) aims to find relevant images based on a textual query, but existing approaches are primarily based on whole-image captions and lack interpretability. Meanwhile, referring expression segmentation (RES) enables precise object localization based on natural language descriptions but is computationally expensive when applied across large image collections. To bridge this gap, we introduce Mask-aware TIR (MaTIR), a new task that unifies TIR and RES, requiring both efficient image search and accurate object segmentation. To address this task, we propose a two-stage framework, comprising a first stage for segmentation-aware image retrieval and a second stage for reranking and object grounding with a multimodal large language model (MLLM). We leverage SAM 2 to generate object masks and Alpha-CLIP to extract region-level embeddings offline at first, enabling effective and scalable online retrieval. Secondly, MLLM is used to refine retrieval rankings and generate bounding boxes, which are matched to segmentation masks. We evaluate our approach on COCO and D$^3$ datasets, demonstrating significant improvements in both retrieval accuracy and segmentation quality over previous methods.