A Sketch+Text Composed Image Retrieval Dataset for Thangka

TL;DR

CIRThan is a new dataset for sketch+text composed image retrieval in Thangka art, highlighting challenges in aligning multimodal inputs with complex, domain-specific imagery.

Contribution

The paper introduces CIRThan, a culturally grounded dataset with hierarchical descriptions for Thangka images, and evaluates existing methods, exposing their limitations in this domain.

Findings

Existing CIR methods struggle with fine-grained, domain-specific Thangka images.

Hierarchical textual descriptions improve semantic understanding in retrieval.

Zero-shot methods perform poorly without in-domain supervision.

Abstract

Composed Image Retrieval (CIR) enables image retrieval by combining multiple query modalities, but existing benchmarks predominantly focus on general-domain imagery and rely on reference images with short textual modifications. As a result, they provide limited support for retrieval scenarios that require fine-grained semantic reasoning, structured visual understanding, and domain-specific knowledge. In this work, we introduce CIRThan, a sketch+text Composed Image Retrieval dataset for Thangka imagery, a culturally grounded and knowledge-specific visual domain characterized by complex structures, dense symbolic elements, and domain-dependent semantic conventions. CIRThan contains 2,287 high-quality Thangka images, each paired with a human-drawn sketch and hierarchical textual descriptions at three semantic levels, enabling composed queries that jointly express structural intent and multi-level semantic specification. We provide standardized data splits, comprehensive dataset analysis, and benchmark evaluations of representative supervised and zero-shot CIR methods. Experimental results reveal that existing CIR approaches, largely developed for general-domain imagery, struggle to effectively align sketch-based abstractions and hierarchical textual semantics with fine-grained Thangka images, particularly without in-domain supervision. We believe CIRThan offers a valuable benchmark for advancing sketch+text CIR, hierarchical semantic modeling, and multimodal retrieval in cultural heritage and other knowledge-specific visual domains. The dataset is publicly available at https://github.com/jinyuxu-whut/CIRThan.