Scholarlink

This study develops a self-supervised Siamese model to predict the relative positions of patches from thoraco-abdominal CT scans, creating an explicit 3D universal coordinate system (UCS) for the human body. The model is evaluated through experiments on body part localization, landmark detection, and longitudinal lesion tracking. Despite the moderate performance of the UCS in these experiments, this model provides a proof-of-concept system for standardized anatomical coordinates. Such a system enables robust identification and comparison of anatomical structures across different individuals and longitudinally.

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

A Radiological-Based Coordinate System for the Human Body: A Proof-of-Concept

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Teresa M. T. Bucho,
  • Thierry N. Boellaard,
  • Mateus Taveira,
  • Zuhir Bodalal,
  • Thi D. L. Nguyen-Kim,
  • Regina Beets-Tan,
  • Stefano Trebeschi

摘要

This study develops a self-supervised Siamese model to predict the relative positions of patches from thoraco-abdominal CT scans, creating an explicit 3D universal coordinate system (UCS) for the human body. The model is evaluated through experiments on body part localization, landmark detection, and longitudinal lesion tracking. Despite the moderate performance of the UCS in these experiments, this model provides a proof-of-concept system for standardized anatomical coordinates. Such a system enables robust identification and comparison of anatomical structures across different individuals and longitudinally.