Metal artifact reduction on cervical CT images by deep residual learning.
Metal artifact reduction on cervical CT images by deep residual learning.
复制标题
通过深度残差学习减少颈部 CT 图像的金属伪影
DOI：
10.1186/s12938-018-0609-y
复制
发表时间：
2018-11-27
影响因子：
3.9
通讯作者：
Zhang Y
中科院分区：
文献类型：
--
作者：
Huang X;Wang J;Tang F;Zhong T;Zhang Y
背景宫颈癌是女性第五大常见癌症，也是全球女性癌症死亡的第三大原因。近距离放射治疗是宫颈癌最有效的治疗方法。对于近距离放射治疗，计算机断层扫描（CT）成像是必要的，因为它传达了可用于剂量规划的组织密度信息。然而，由近距离放射治疗施源器引起的金属伪影仍然是用于图像引导程序的图像数据的自动处理或精确剂量计算的挑战。因此，开发一个有效的金属伪影减少（MAR）算法在宫颈CT images.MethodsA新的残差学习方法的基础上卷积神经网络（RL-ARCNN），提出了减少宫颈CT图像中的金属伪影。对于MAR，在第一步中通过模拟各种金属伪影生成数据集，该数据集将用于训练CNN。该数据集包括伪影插入图像、无伪影图像和伪影残留图像。从数据集中提取大量图像块，用于基于CNN（RL-ARCNN）的深度残差学习伪影减少训练。之后，训练的模型可以用于MAR对宫颈CT images.ResultsThe所提出的方法提供了一个很好的MAR结果与PSNR为38.09的模拟伪影图像的测试集。残差学习的PSNR（38.09）高于普通学习的PSNR（37.79），这表明基于CNN的残差图像实现了有利的伪影减少。此外，对于512 × 512的图像，平均去除伪影的时间小于1 s。结论RL-ARCNN表明，CNN的剩余学习显着减少金属伪影，提高关键结构的可视化和放射肿瘤学家在目标划定的信心。金属伪影被有效地消除，没有正弦图数据和复杂的后处理过程。
BackgroundCervical cancer is the fifth most common cancer among women, which is the third leading cause of cancer death in women worldwide. Brachytherapy is the most effective treatment for cervical cancer. For brachytherapy, computed tomography (CT) imaging is necessary since it conveys tissue density information which can be used for dose planning. However, the metal artifacts caused by brachytherapy applicators remain a challenge for the automatic processing of image data for image-guided procedures or accurate dose calculations. Therefore, developing an effective metal artifact reduction (MAR) algorithm in cervical CT images is of high demand.MethodsA novel residual learning method based on convolutional neural network (RL-ARCNN) is proposed to reduce metal artifacts in cervical CT images. For MAR, a dataset is generated by simulating various metal artifacts in the first step, which will be applied to train the CNN. This dataset includes artifact-insert, artifact-free, and artifact-residual images. Numerous image patches are extracted from the dataset for training on deep residual learning artifact reduction based on CNN (RL-ARCNN). Afterwards, the trained model can be used for MAR on cervical CT images.ResultsThe proposed method provides a good MAR result with a PSNR of 38.09 on the test set of simulated artifact images. The PSNR of residual learning (38.09) is higher than that of ordinary learning (37.79) which shows that CNN-based residual images achieve favorable artifact reduction. Moreover, for a 512 × 512 image, the average removal artifact time is less than 1 s.ConclusionsThe RL-ARCNN indicates that residual learning of CNN remarkably reduces metal artifacts and improves critical structure visualization and confidence of radiation oncologists in target delineation. Metal artifacts are eliminated efficiently free of sinogram data and complicated post-processing procedure.
登录
查看更多内容
影响因子：
6.7
作者：
Aissa, Joel;Boos, Johannes;Thomas, Christoph
通讯作者：
Thomas, Christoph
影响因子：
19.7
作者：
Boas, F. Edward;Fleischmann, Dominik
通讯作者：
Fleischmann, Dominik
影响因子：
19.7
作者：
KALENDER, WA;HEBEL, R;EBERSBERGER, J
通讯作者：
EBERSBERGER, J
通过卷积神经网络进行低剂量 CT
影响因子：
3.4
作者：
Chen, Hu;Zhang, Yi;Wang, Ge
通讯作者：
Wang, Ge
DOI：
10.1016/s0360-3016(01)01664-9
发表时间：
2001-09-01
影响因子：
7
作者：
Portelance, L;Chao, KSC;Low, D
通讯作者：
Low, D