Department of Diagnostic Imaging, St Jude Children's Research Hospital, Memphis, TN 38105, United States of America. Radiology Department, Cincinnati Children's Hospital, Cincinnati, OH 45229, United States of America.
Phys Med Biol. 2018 Jun 27;63(13):135009. doi: 10.1088/1361-6560/aac944.
To develop a consistent, fully-automated classifier for all tissues within the trunk and to more accurately discriminate between tissues (such as bone) and contrast medium with overlapping high CT numbers. Twenty-eight contrast enhanced NCAP (neck-chest-abdomen-pelvis) CT scans (four adult and three pediatric patients) were used to train and test a tissue classification pipeline. The classifier output consisted of six tissue classes: lung, fat, muscle, solid organ, blood/bowel contrast and bone. The input features for training were selected from 28 2D image filters and 12 3D filters, and one hand crafted spatial feature. To improve differentiation between tissue and blood/bowel contrast classification, 70 additional CT images were manually classified. Two different training data sets consisting of manually classified tissues from different locations in body were used to train the models. Training used the random forest algorithm in WEKA (Waikato Environment for Knowledge Analysis); the number of trees was optimized for best out-of-bag error. Automated classification accuracy was compared with manual classification by calculating dice similarity coefficient (DSC). Model performance was tested on 21 manually classified slices (two adult and one pediatric patient). The overall DSC at image locations represented in the training dataset were-lung: 0.98, fat: 0.90, muscle: 0.85, solid organ: 0.75, blood/contrast: 0.82, and bone: 0.90. The overall DSC for slice locations that were not represented in the training dataset were-lung: 0.97, fat: 0.89, muscle: 0.76, solid organ: 0.79, blood: 0.56, and bone: 0.74. Analyzing the classification maps for the entire scan volume revealed that except for misclassifications in the trabecular bone region of the spinal column, and solid organ and blood/contrast interfaces within the abdomen, the results were acceptable. A fully-automated whole-body tissue classifier for adult and pediatric contrast-enhanced CT using random forest algorithm and intensity-based image filters was developed.
为了开发一种用于躯干所有组织的一致的全自动分类器，并更准确地区分组织（如骨骼）和具有重叠高 CT 数的对比剂。使用 28 例增强型 NCAP（颈胸腹部骨盆）CT 扫描（4 例成人和 3 例儿科患者）来训练和测试组织分类管道。分类器输出包括六个组织类别：肺、脂肪、肌肉、实体器官、血液/肠腔对比和骨骼。用于训练的输入特征选自 28 个 2D 图像滤波器和 12 个 3D 滤波器，以及一个手工制作的空间特征。为了改善组织和血液/肠腔对比分类之间的区分，手动分类了 70 个额外的 CT 图像。使用来自身体不同部位的手动分类组织的两个不同的训练数据集来训练模型。训练使用了 WEKA（Waikato Environment for Knowledge Analysis）中的随机森林算法；对树木的数量进行了优化，以获得最佳的袋外误差。通过计算骰子相似系数（DSC）比较自动分类精度与手动分类。在 21 个手动分类切片（2 个成人和 1 个儿科患者）上测试了模型性能。在训练数据集表示的图像位置的总体 DSC 分别为-肺：0.98、脂肪：0.90、肌肉：0.85、实体器官：0.75、血液/对比：0.82 和骨骼：0.90。在未在训练数据集中表示的切片位置的总体 DSC 分别为-肺：0.97、脂肪：0.89、肌肉：0.76、实体器官：0.79、血液：0.56 和骨骼：0.74。分析整个扫描体积的分类图表明，除了脊柱小梁骨区域、腹部实体器官和血液/对比界面的误分类外，结果是可以接受的。使用随机森林算法和基于强度的图像滤波器为成人和儿科增强 CT 开发了一种用于全身组织的全自动分类器。
