HRVVS: A High-resolution Video Vasculature Segmentation Network via Hierarchical Autoregressive Residual Priors

2025-07-30

AI Summary

High-precision segmentation of hepatic vasculature in liver resection videos remains challenging due to severe scarcity of annotated data and anatomical complexity. To address this, we introduce the first large-scale surgical video dataset—comprising 35 long videos with frame-level annotations across 11,442 frames—and propose HRVVS, a novel network architecture. HRVVS pioneers the integration of a pre-trained vision autoregressive model as a multi-level prior-embedding encoder to mitigate information degradation during downsampling; incorporates a dynamic memory decoder to suppress redundant feature propagation; and unifies hierarchical autoregressive residual modeling, multi-view segmentation, and high-resolution feature preservation. Evaluated on a dedicated surgical video benchmark, HRVVS achieves state-of-the-art performance, delivering substantial improvements in both segmentation accuracy and fine-grained vascular structure preservation.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Large Multimodal Models (LMMs)
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

The segmentation of the hepatic vasculature in surgical videos holds substantial clinical significance in the context of hepatectomy procedures. However, owing to the dearth of an appropriate dataset and the inherently complex task characteristics, few researches have been reported in this domain. To address this issue, we first introduce a high quality frame-by-frame annotated hepatic vasculature dataset containing 35 long hepatectomy videos and 11442 high-resolution frames. On this basis, we propose a novel high-resolution video vasculature segmentation network, dubbed as HRVVS. We innovatively embed a pretrained visual autoregressive modeling (VAR) model into different layers of the hierarchical encoder as prior information to reduce the information degradation generated during the downsampling process. In addition, we designed a dynamic memory decoder on a multi-view segmentation network to minimize the transmission of redundant information while preserving more details between frames. Extensive experiments on surgical video datasets demonstrate that our proposed HRVVS significantly outperforms the state-of-the-art methods. The source code and dataset will be publicly available at href{https://github.com/scott-yjyang/xx}{https://github.com/scott-yjyang/HRVVS}.

Problem

• Segment hepatic vasculature in surgical videos for hepatectomy
• Lack of annotated dataset for video vasculature segmentation
• Address information degradation in downsampling during segmentation

Innovation

• Hierarchical encoder with autoregressive residual priors
• Dynamic memory decoder for multi-view segmentation
• High-resolution video vasculature segmentation network