BREATH-VL: Vision-Language-Guided 6-DoF Bronchoscopy Localization via Semantic-Geometric Fusion

TL;DR

This paper introduces BREATH-VL, a hybrid vision-language framework for accurate 6-DoF bronchoscopy localization, leveraging a new in-vivo dataset and semantic-geometric fusion to improve accuracy and robustness in complex airway navigation.

Contribution

The paper presents BREATH-VL, the first in-vivo endoscopic localization dataset and a hybrid framework combining vision-language cues with geometric registration for enhanced 6-DoF pose estimation.

Findings

Reduces translational error by 25.5% compared to state-of-the-art methods.

Demonstrates robust semantic localization in challenging surgical scenes.

Achieves competitive computational latency with improved accuracy.

Abstract

Vision-language models (VLMs) have recently shown remarkable performance in navigation and localization tasks by leveraging large-scale pretraining for semantic understanding. However, applying VLMs to 6-DoF endoscopic camera localization presents several challenges: 1) the lack of large-scale, high-quality, densely annotated, and localization-oriented vision-language datasets in real-world medical settings; 2) limited capability for fine-grained pose regression; and 3) high computational latency when extracting temporal features from past frames. To address these issues, we first construct BREATH dataset, the largest in-vivo endoscopic localization dataset to date, collected in the complex human airway. Building on this dataset, we propose BREATH-VL, a hybrid framework that integrates semantic cues from VLMs with geometric information from vision-based registration methods for accurate 6-DoF pose estimation. Our motivation lies in the complementary strengths of both approaches: VLMs offer generalizable semantic understanding, while registration methods provide precise geometric alignment. To further enhance the VLM's ability to capture temporal context, we introduce a lightweight context-learning mechanism that encodes motion history as linguistic prompts, enabling efficient temporal reasoning without expensive video-level computation. Extensive experiments demonstrate that the vision-language module delivers robust semantic localization in challenging surgical scenes. Building on this, our BREATH-VL outperforms state-of-the-art vision-only localization methods in both accuracy and generalization, reducing translational error by 25.5% compared with the best-performing baseline, while achieving competitive computational latency.