Scholarlink

Calcium scoring is a fundamental technique for assessing the extent of calcification in the cardiovascular system, which can be used for diagnostic purposes and predicting the occurrence of major adverse cardiac events (MACE). Although it is a standard procedure, the process is quite challenging and ambiguous, often resulting in varying results between different specialists. In addition, manual identification of coronary calcification is time-consuming and impractical on a large scale, thus increasing the demand for automated methods. This work aims to evaluate various deep learning models and forms of data processing (segmenting per 2 d slice vs using 3 d sliding window) used for the problem of fully-automatic plaque segmentation from non-contrast cardiac CT images with subsequent plaque volume and Agatson score estimation.

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

The Comparison of 2D and 3D Based Models for the Problem of Plaque Segmentation and Coronary Artery Calcium Scoring on Non-contrast Cardiac CT Imaging

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jakub Chojnacki,
  • Miłosz Gajowczyk,
  • Kinga Teklak,
  • Tomasz Konopczyński,
  • Maciej Zamorski

摘要

Calcium scoring is a fundamental technique for assessing the extent of calcification in the cardiovascular system, which can be used for diagnostic purposes and predicting the occurrence of major adverse cardiac events (MACE). Although it is a standard procedure, the process is quite challenging and ambiguous, often resulting in varying results between different specialists. In addition, manual identification of coronary calcification is time-consuming and impractical on a large scale, thus increasing the demand for automated methods. This work aims to evaluate various deep learning models and forms of data processing (segmenting per 2 d slice vs using 3 d sliding window) used for the problem of fully-automatic plaque segmentation from non-contrast cardiac CT images with subsequent plaque volume and Agatson score estimation.