Scholarlink

<p>Subarachnoid hemorrhages, often caused by ruptured cerebral aneurysms, require precise vessel segmentation for early intervention and surgical planning. Volumetric segmentation of cerebral vasculature is essential for stroke screening and treatment response assessment. However, the small size and complex topology of cerebral vessels pose significant challenges for clinically reliable segmentation. In this paper, we propose a novel dilated residual-based network for segmenting the major cerebral vessels. The major cerebral vessels provide high contextual information of location of aneurysms. Anatomical information of the location of cerebral aneurysm remnants is used for better segmentation and lightweight network development. An extensive quantitative and visual assessment has been done with state-of-the-art networks for cerebral vessel segmentation. Our proposed method demonstrated promising result with dice score of 0.94 on in-house Aneurysm Database. Furthermore, with the help of neuro-interventional radiologists, we have analyzed the relevance of major cerebral vessels segmentation for aneurysm quantification in streamlining endovascular surgical planning as the method attains higher level of accuracy and maintains consistency in preserving vascular pathology. A novel, robust, cerebral vessel segmentation method was proposed. The method provides the relevance of vessel segmentation, paving the way for improved diagnostic accuracy and clinical decision-making in intracranial aneurysms.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Computationally efficient dilated residual networks for segmentation of major cerebral vessels in MRA

  • 期刊论文

  • 发表日期：

    2025年9月3日

查看全文
  • Subhash Chandra Pal,
  • Chirag Kamal Ahuja,
  • Dimitrios Toumpanakis,
  • Johan Wikstrom,
  • Robin Strand,
  • Ashis Kumar Dhara

摘要

Subarachnoid hemorrhages, often caused by ruptured cerebral aneurysms, require precise vessel segmentation for early intervention and surgical planning. Volumetric segmentation of cerebral vasculature is essential for stroke screening and treatment response assessment. However, the small size and complex topology of cerebral vessels pose significant challenges for clinically reliable segmentation. In this paper, we propose a novel dilated residual-based network for segmenting the major cerebral vessels. The major cerebral vessels provide high contextual information of location of aneurysms. Anatomical information of the location of cerebral aneurysm remnants is used for better segmentation and lightweight network development. An extensive quantitative and visual assessment has been done with state-of-the-art networks for cerebral vessel segmentation. Our proposed method demonstrated promising result with dice score of 0.94 on in-house Aneurysm Database. Furthermore, with the help of neuro-interventional radiologists, we have analyzed the relevance of major cerebral vessels segmentation for aneurysm quantification in streamlining endovascular surgical planning as the method attains higher level of accuracy and maintains consistency in preserving vascular pathology. A novel, robust, cerebral vessel segmentation method was proposed. The method provides the relevance of vessel segmentation, paving the way for improved diagnostic accuracy and clinical decision-making in intracranial aneurysms.