NCI-ISBI 2013 Challenge: Automated Segmentation of Prostate Structures

This data set was created for use in the <a href="https://wiki.cancerimagingarchive.net/display/Public/Challenge+competitions">NCI-ISBI 2013 Challenge - Automated Segmentation of Prostate Structures</a>.  The challenge data set was divided into 3 parts including training, leaderboard and test data sets.  This allowed participants to prepare their algorithms and test their results prior to submitting to a final test phase for selecting the winner.Image data were selected from <a href="https://cancerimagingarchive.net/collection/prostate-diagnosis/" target="_blank" rel="noopener">PROSTATE-DIAGNOSIS</a> and <a href="https://cancerimagingarchive.net/collection/prostate-3t/" target="_blank" rel="noopener">Prostate-3T</a> collections on TCIA.  Cases consist of axial scans with half obtained at 1.5 T (Philips Achieva) with an endo-rectal receiver coil (fromBostonMedicalCenter) and the other half at 3T (Siemens TIM) with a surface coil (from Radboud University Nijmegen Medical Centre [RUNMC],Netherlands). They were acquired as T2-weighted MR axial pulse sequences with either 4 mm thick slices at 3T or 3 mm thick at 1.5T. Each file contains nearly all DICOM acquisition parameters except tags that specifically identify Private Health Information.Each case has had central gland (CG) and peripheral zone (PZ) outlines marked by Drs. Nicolas Bloch (Boston University School of Medicine) and Mirabela Rusu (Case Western University) or Drs. Henkjan Huisman, Geert Litjens, or Jurgen Futterer at RUNMC Netherlands.

本数据集旨在应用于2013年NCI-ISBI挑战赛——前列腺结构自动分割竞赛（<a href="https://wiki.cancerimagingarchive.net/display/Public/Challenge+competitions">NCI-ISBI 2013 Challenge - Automated Segmentation of Prostate Structures</a>）。挑战数据集分为训练集、排行榜集和测试集三部分，以便参赛者能够准备算法并测试结果，在提交最终测试阶段以选拔优胜者。图像数据选自TCIA平台上的<em>PROSTATE-DIAGNOSIS</em>和<em>Prostate-3T</em>两个系列，其中一半图像采用1.5T（Philips Achieva）的直肠内接收线圈（来自波士顿医疗中心）获取，另一半图像采用3T（Siemens TIM）的表面线圈（来自荷兰奈梅亨拉德布罗伊尔特大学医疗中心[Radboud University Nijmegen Medical Centre, RUNMC]）获取。所有图像均以T2加权磁共振轴位脉冲序列进行采集，切片厚度分别为3T下的4毫米和1.5T下的3毫米。每个文件包含几乎所有的DICOM采集参数，但排除了用于识别私人健康信息的特定标签。每个病例均由Boston University School of Medicine的Nicolas Bloch博士和Case Western University的Mirabela Rusu博士，或在RUNMC荷兰的Henkjan Huisman、Geert Litjens或Jurgen Futterer博士标记了中心腺体（CG）和周围带（PZ）的轮廓。

搜集汇总

背景与挑战

背景概述

该数据集是为NCI-ISBI 2013挑战赛设计的，专门用于前列腺结构的自动分割任务。数据来源于TCIA的两个前列腺MRI集合，包含1.5T和3T的T2加权轴向扫描图像，并已由医学专家标注了中央腺体和周围区的轮廓。数据集分为训练、排行榜和测试三部分，支持算法开发和公平评估。