Scholarlink

Introduction <p>Previous algorithms for measuring laser alignment in the American College of Radiology (ACR) computed tomography (CT) phantom sometimes failed to detect ball bearings (BBs) due to the presence of aluminum wires, and it failed if all BBs in the image were not clearly visible. This study proposes an improved algorithm for automated measurement of laser alignment in the ACR CT phantom.</p> Methods <p>The improved algorithm employed a masking approach to remove central objects, leaving only the BBs at the phantom’s edges. Since the BBs often do not appear simultaneously in all four positions, the algorithm accounted for this by rotating the image based on their coordinates, allowing measurements to proceed even if only two or three BBs were visible. Evaluation of the algorithm was conducted using modules #1 and #4 of the ACR CT phantom scanned with slice thicknesses from 1.5 to 10&#xa0;mm.</p> Results <p>The improved algorithm successfully performed measurements in both modules for all slices, whereas a previous algorithm failed to accurately capture objects in module #1 at a slice thickness of 1.5&#xa0;mm. Additionally, the new algorithm enabled measurements on some images where only two or three BBs were visible.</p> Conclusion <p>The new developed algorithm is more robust at measuring laser alignment in the ACR CT phantom. It is accurate in easy to be implemented in the clinical setting.</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

An improved method for automated measurement of laser alignment using the ACR CT phantom

  • 期刊论文

  • 发表日期：

    2025年9月20日

查看全文
  • Choirul Anam,
  • Riska Amilia,
  • Ariij Naufal,
  • Heri Sutanto,
  • Kusworo Adi,
  • Geoff Dougherty

摘要

Introduction

Previous algorithms for measuring laser alignment in the American College of Radiology (ACR) computed tomography (CT) phantom sometimes failed to detect ball bearings (BBs) due to the presence of aluminum wires, and it failed if all BBs in the image were not clearly visible. This study proposes an improved algorithm for automated measurement of laser alignment in the ACR CT phantom.

Methods

The improved algorithm employed a masking approach to remove central objects, leaving only the BBs at the phantom’s edges. Since the BBs often do not appear simultaneously in all four positions, the algorithm accounted for this by rotating the image based on their coordinates, allowing measurements to proceed even if only two or three BBs were visible. Evaluation of the algorithm was conducted using modules #1 and #4 of the ACR CT phantom scanned with slice thicknesses from 1.5 to 10 mm.

Results

The improved algorithm successfully performed measurements in both modules for all slices, whereas a previous algorithm failed to accurately capture objects in module #1 at a slice thickness of 1.5 mm. Additionally, the new algorithm enabled measurements on some images where only two or three BBs were visible.

Conclusion

The new developed algorithm is more robust at measuring laser alignment in the ACR CT phantom. It is accurate in easy to be implemented in the clinical setting.