Saved in:
Bibliographic Details
Main Authors: Liu, Yaoyu, Zhang, Minghui, You, Xin, Zhang, Hanxiao, Gu, Yun
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.09217
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908883098796032
author Liu, Yaoyu
Zhang, Minghui
You, Xin
Zhang, Hanxiao
Gu, Yun
author_facet Liu, Yaoyu
Zhang, Minghui
You, Xin
Zhang, Hanxiao
Gu, Yun
contents Modeling medical vessel-like anatomy is challenging due to its intricate topology and sensitivity to dataset shifts. Consequently, task-specific models often suffer from topological inconsistencies, including artificial disconnections and spurious merges. Motivated by the promise of multimodal large language models (MLLMs) for zero-shot generalization, we propose TubeMLLM, a unified foundation model that couples structured understanding with controllable generation for medical vessel-like anatomy. By integrating topological priors through explicit natural language prompting and aligning them with visual representations in a shared-attention architecture, TubeMLLM significantly enhances topology-aware perception. Furthermore, we construct TubeMData, a pionner multimodal benchmark comprising comprehensive topology-centric tasks, and introduce an adaptive loss weighting strategy to emphasize topology-critical regions during training. Extensive experiments on fifteen diverse datasets demonstrate our superiority. Quantitatively, TubeMLLM achieves state-of-the-art out-of-distribution performance, substantially reducing global topological discrepancies on color fundus photography (decreasing the $β_{0}$ number error from 37.42 to 8.58 compared to baselines). Notably, TubeMLLM exhibits exceptional zero-shot cross-modality transferring ability on unseen X-ray angiography, achieving a Dice score of 67.50% while significantly reducing the $β_{0}$ error to 1.21. TubeMLLM also maintains robustness against degradations such as blur, noise, and low resolution. Furthermore, in topology-aware understanding tasks, the model achieves 97.38% accuracy in evaluating mask topological quality, significantly outperforming standard vision-language baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2603_09217
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle TubeMLLM: A Foundation Model for Topology Knowledge Exploration in Vessel-like Anatomy
Liu, Yaoyu
Zhang, Minghui
You, Xin
Zhang, Hanxiao
Gu, Yun
Computer Vision and Pattern Recognition
Modeling medical vessel-like anatomy is challenging due to its intricate topology and sensitivity to dataset shifts. Consequently, task-specific models often suffer from topological inconsistencies, including artificial disconnections and spurious merges. Motivated by the promise of multimodal large language models (MLLMs) for zero-shot generalization, we propose TubeMLLM, a unified foundation model that couples structured understanding with controllable generation for medical vessel-like anatomy. By integrating topological priors through explicit natural language prompting and aligning them with visual representations in a shared-attention architecture, TubeMLLM significantly enhances topology-aware perception. Furthermore, we construct TubeMData, a pionner multimodal benchmark comprising comprehensive topology-centric tasks, and introduce an adaptive loss weighting strategy to emphasize topology-critical regions during training. Extensive experiments on fifteen diverse datasets demonstrate our superiority. Quantitatively, TubeMLLM achieves state-of-the-art out-of-distribution performance, substantially reducing global topological discrepancies on color fundus photography (decreasing the $β_{0}$ number error from 37.42 to 8.58 compared to baselines). Notably, TubeMLLM exhibits exceptional zero-shot cross-modality transferring ability on unseen X-ray angiography, achieving a Dice score of 67.50% while significantly reducing the $β_{0}$ error to 1.21. TubeMLLM also maintains robustness against degradations such as blur, noise, and low resolution. Furthermore, in topology-aware understanding tasks, the model achieves 97.38% accuracy in evaluating mask topological quality, significantly outperforming standard vision-language baselines.
title TubeMLLM: A Foundation Model for Topology Knowledge Exploration in Vessel-like Anatomy
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.09217