Гулака Правин (KR),Коробченко Дмитрий Александрович (RU),Сиротенко Михаил Юрьевич (RU),Данилевич Алексей Брониславович (RU),Рычагов Михаил Николаевич (RU)
申请号：
RU2013111932/08
公开号：
RU2013111932A
申请日:
2013.03.18
申请国别(地区):
RU
年份:
2014
代理人:
摘要：
1. A method for automatically registering anatomical points in voluminous medical images, involving operations such as obtaining a three-dimensional medical image, determining a plurality of search points, extracting features of said points, generating a plurality of candidates for the desired anatomical points, filtering these candidates, displaying the final positions of the desired anatomical points , characterized in that: - the specified definition of the set of search points is produced by setting the search grid then a check inside a three-dimensional image using a statistical atlas； the specified statistical atlas is attached to the bounding box corresponding to the studied part of the body inside the three-dimensional image； - the indicated feature extraction of points is done by selecting the surrounding context for each such point and applying a multilayer feature extraction algorithm to it； - the specified formation of a set of candidates is performed by calculating the measure, qualities for each search point and for each type of desired anatomical points using the corresponding algorithm output feature attraction； - the specified filtering of candidates is performed using a statistical atlas and thresholds calculated in advance for the values of the quality measure； - the indicated conclusion of the final positions is made by sorting all the candidates remaining after filtering within each type of the desired anatomical points according to their quality and the candidates with the highest quality measure values or the required number of candidates with the highest quality measure values1. Способ автоматической регистрации анатомических точек вобъёмных медицинских изображениях, предусматривающий выполнение таких операций, как получение трёхмерного медицинского изображения, определение множества поисковых точек, извлечение признаков указанных точек, формирование множества кандидатов на искомые анатомические точки, фильтрация указанных кандидатов, вывод ф
