Hyperbolic Hierarchical Alignment Reasoning Network for Text-3D Retrieval

TL;DR

This paper introduces H$^{2}$ARN, a hyperbolic space-based model for text-3D retrieval that effectively preserves hierarchical structures and reduces redundancy, achieving improved discriminative ability on a new large-scale benchmark.

Contribution

The paper proposes a novel hyperbolic embedding approach with hierarchical and contribution-aware modules for text-3D retrieval, addressing hierarchy collapse and saliency dilution issues.

Findings

H$^{2}$ARN outperforms existing methods on the T3DR-HIT v2 benchmark.

The hyperbolic space preserves hierarchical relationships more effectively.

Contribution-aware aggregation enhances discriminative feature selection.

Abstract

With the daily influx of 3D data on the internet, text-3D retrieval has gained increasing attention. However, current methods face two major challenges: Hierarchy Representation Collapse (HRC) and Redundancy-Induced Saliency Dilution (RISD). HRC compresses abstract-to-specific and whole-to-part hierarchies in Euclidean embeddings, while RISD averages noisy fragments, obscuring critical semantic cues and diminishing the model's ability to distinguish hard negatives. To address these challenges, we introduce the Hyperbolic Hierarchical Alignment Reasoning Network (H$^{2}$ARN) for text-3D retrieval. H$^{2}$ARN embeds both text and 3D data in a Lorentz-model hyperbolic space, where exponential volume growth inherently preserves hierarchical distances. A hierarchical ordering loss constructs a shrinking entailment cone around each text vector, ensuring that the matched 3D instance falls within the cone, while an instance-level contrastive loss jointly enforces separation from non-matching samples. To tackle RISD, we propose a contribution-aware hyperbolic aggregation module that leverages Lorentzian distance to assess the relevance of each local feature and applies contribution-weighted aggregation guided by hyperbolic geometry, enhancing discriminative regions while suppressing redundancy without additional supervision. We also release the expanded T3DR-HIT v2 benchmark, which contains 8,935 text-to-3D pairs, 2.6 times the original size, covering both fine-grained cultural artefacts and complex indoor scenes. Our codes are available at https://github.com/liwrui/H2ARN.