A Deep Learning Approach for Rapid and Generalizable Denoising of Photon-Counting Micro-CT Images.

A Deep Learning Approach for Rapid and Generalizable Denoising of Photon-Counting Micro-CT Images.
复制标题

DOI:
10.3390/tomography9040102
复制
发表时间:
2023-07-02
期刊:
Tomography (Ann Arbor, Mich.)
影响因子:
--
通讯作者:
Badea CT
Badea CT
中科院分区:
其他
文献类型:
--
作者:
Nadkarni R;Clark DP;Allphin AJ;Badea CT

文献摘要

参考文献

相似文献

光子计数CT(PCCT)在光谱成像和材料分解方面功能强大,但会产生噪声加权滤波反投影(WFBP)重建。虽然迭代重建有效地去除了这些图像的噪声,但它需要大量的计算时间。为了克服这一局限性,我们提出了一种深度学习模型UnetU,该模型能够快速地估计wFBP的迭代重构。UnetU利用具有自定义损失函数和wFBP变换的2D U-Net卷积神经网络(CNN),在各种光子计数探测器(PCD)能量阈值设置中促进准确的材料分解。UnetU在测试集重构中的均方根误差(RMSE)和它们各自的矩阵求逆材料分解方面优于多能量非局部平均(ME NLM)和传统的去噪CNN UnetwFBP。重建和材料分解领域的定性结果表明,UnetU是迭代重建的最佳逼近。在具有不同欠采样因子的高剂量体外扫描重建中,UnetU一致地给出了比ME NLM和UnetwFBP更高的结构相似性(SSIM)和峰值信噪比(PSNR)。这项研究展示了UnetU作为一种快速(即,比迭代重建快15倍)和可推广的PCCT去噪方法的潜力,为推进临床前PCCT研究提供了希望。
Photon-counting CT (PCCT) is powerful for spectral imaging and material decomposition but produces noisy weighted filtered backprojection (wFBP) reconstructions. Although iterative reconstruction effectively denoises these images, it requires extensive computation time. To overcome this limitation, we propose a deep learning (DL) model, UnetU, which quickly estimates iterative reconstruction from wFBP. Utilizing a 2D U-net convolutional neural network (CNN) with a custom loss function and transformation of wFBP, UnetU promotes accurate material decomposition across various photon-counting detector (PCD) energy threshold settings. UnetU outperformed multi-energy non-local means (ME NLM) and a conventional denoising CNN called UnetwFBP in terms of root mean square error (RMSE) in test set reconstructions and their respective matrix inversion material decompositions. Qualitative results in reconstruction and material decomposition domains revealed that UnetU is the best approximation of iterative reconstruction. In reconstructions with varying undersampling factors from a high dose ex vivo scan, UnetU consistently gave higher structural similarity (SSIM) and peak signal-to-noise ratio (PSNR) to the fully sampled iterative reconstruction than ME NLM and UnetwFBP. This research demonstrates UnetU’s potential as a fast (i.e., 15 times faster than iterative reconstruction) and generalizable approach for PCCT denoising, holding promise for advancing preclinical PCCT research.
DOI: 10.1088/0031-9155/61/18/6707
发表时间: 2016-09-21
影响因子: 3.5
作者:
Yu Z;Leng S;Li Z;McCollough CH
通讯作者: McCollough CH
DOI: 10.1088/1361-6560/ac7d34
发表时间: 2022-07-18
影响因子: 3.5
作者:
Nadkarni, Rohan;Allphin, Alex;Clark, Darin P.;Badea, Cristian T.
通讯作者: Badea, Cristian T.
DOI: 10.1109/tip.2017.2713099
发表时间: 2017-09-01
影响因子: 10.6
作者:
Jin, Kyong Hwan;McCann, Michael T.;Unser, Michael
通讯作者: Unser, Michael
DOI: 10.1118/1.4820371
发表时间: 2013-10-01
期刊: MEDICAL PHYSICS
影响因子: 3.8
作者:
Taguchi, Katsuyuki;Iwanczyk, Jan S.
通讯作者: Iwanczyk, Jan S.
DOI: 10.1088/1361-6560/aba8b2
发表时间: 2020-10-21
影响因子: 3.5
作者:
Holbrook MD;Clark DP;Badea CT
通讯作者: Badea CT