首页> 外国专利> Method and apparatus for selecting a subset of atlases from a set of candidate atlases

Method and apparatus for selecting a subset of atlases from a set of candidate atlases

机译:从一组候选地图集中选择地图集的子集的方法和设备

摘要

An auto-contouring performance measure (e.g. DICE coefficient, Jaccard index, Hausdorff distance, mean distance to agreement) is computed (335) for each candidate atlas n within the set of N candidate atlases (110) in relation to each training atlas t within a subset of a set of T training atlases (330). For each pair combination of n and t for which a measure has been computed, a score (e.g. relative weighting, absolute weighting, editing time prediction value) is derived (360) based on the computed measures. The derived scores are aggregated (365) for each candidate n within the set N for each pair combination comprising that candidate atlas n. A subset of M atlases is selected (370) from the set of N atlases based on the aggregated scores (e.g. those with the most favourable scores) which can be used by medical image auto-contouring systems. Auto-contouring performance measures may be computed for each of a plurality of structures of each atlas n within the set N in relation to each training atlas t, and a score may be derived for each of the plurality of structures, e.g. partly based on a structure weighting value. An atlas is a medical image with previously delineated structure(s).
机译:针对在N个候选地图集(110)中的每个候选地图集n中的每个候选地图集n,计算(335)自动轮廓性能度量(例如,DICE系数,Jaccard指数,Hausdorff距离,平均一致距离) T训练图集的子集(330)。对于已经针对其计算了度量的n和t的每对组合,基于所计算的度量来导出分数(例如,相对加权,绝对加权,编辑时间预测值)(360)。对于包括该候选图集n的每个对组合,针对集合N内的每个候选n汇总导出的分数(365)。基于医学图像自动轮廓系统可以使用的合计得分(例如,具有最有利得分的得分),从N个地图集的集合中选择M个地图集的子集(370)。可以针对与每个训练图集t有关的集合N内的每个图集n的多个结构中的每一个来计算自动轮廓性能度量,并且可以针对多个结构中的每一个来得出分数,例如,针对每个训练图谱t。部分基于结构权重值。地图集是具有先前描述的结构的医学图像。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

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