首页> 外文学位 >Digital processing of neurologic images.
【24h】

Digital processing of neurologic images.

机译:神经影像的数字处理。

获取原文
获取原文并翻译 | 示例

摘要

Digital processing can be used to greatly enhance the utility of medical images. This document includes analysis and development of multiple digital image processing methods, both individually and in aggregate, for neurologic image applications. We investigated parameter optimization and novel extension of the nonlocal means denoising algorithm, showing that using a very small search window can result in increased denoising performance with proper adjustment of the other parameters, and that the nonlocal property can be exploited to use thick slice reconstructions in denoising thin slice reconstructions. After developing a whole brain segmentation algorithm for a valuable set of magnetic resonance (MR) exams of multiple sclerosis patients with biopsy-proven immunopattern classifications, we present analysis showing no significant tissue- and acquisition-specific MR signal intensity differences between the immunopatterns. Finally, we discovered clinically significant differences in measurement output from FDA-approved software used for computing cerebral blood volume maps from dynamic susceptibility contrast MR images. The potential clinical impact of these differences is shown to be disagreement between the software 20-50% of the time regarding classification of post-radiation treatment glioblastoma multiforme patients as having progressing or pseudoprogressing tumor. Additionally, the results indicate that 3.0 T MR acquisitions provide more useful perfusion information for identification of tumor pseudoprogression than 1.5 T acquisitions.
机译:可以使用数字处理来大大增强医学图像的实用性。本文档包括针对神经图像应用的多种数字图像处理方法的分析和开发,无论是单独的还是整体的。我们研究了参数优化和非局部均值去噪算法的新型扩展,结果表明,使用很小的搜索窗口可以通过适当调整其他参数来提高去噪性能,并且可以利用非局部属性在粗加工中使用厚切片重建去噪薄片重建。在开发出一套完整的脑部分割算法,以对多发性硬化症患者进行活检验证的免疫模式分类后,进行有价值的磁共振(MR)检查,然后进行分析,表明免疫模式之间没有明显的组织特异性和采集特异性MR信号强度差异。最后,我们发现了来自FDA批准的用于从动态磁化率对比MR图像计算脑血量图的软件的测量输出中的临床显着差异。这些差异的潜在临床影响显示出,在20-50%的时间中,对于放射治疗后多形性胶质母细胞瘤患者是否具有进展性或假性肿瘤分类,软件之间存在分歧。此外,结果表明,与1.5 T采集相比,3.0 T MR采集可提供更多有用的灌注信息来鉴定肿瘤的假进展。

著录项

  • 作者

    Kelm, Zachary Samuel.;

  • 作者单位

    College of Medicine - Mayo Clinic.;

  • 授予单位 College of Medicine - Mayo Clinic.;
  • 学科 Medical imaging.;Biomedical engineering.
  • 学位 Ph.D.
  • 年度 2015
  • 页码 103 p.
  • 总页数 103
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号