Generative Adversarial Networks for Noise Reduction in Low-Dose CT

Jelmer M. Wolterink; Tim Leiner; Max A. Viergever; Ivana I?gum

<track id="lqepq"><em id="lqepq"></em></track>

首頁> 外文期刊>IEEE Transactions on Medical Imaging >文獻詳情

【24h】

Generative Adversarial Networks for Noise Reduction in Low-Dose CT

機譯：用于小劑量CT的降噪的生成對抗網絡

獲取原文

獲取原文并翻譯 | 示例

頁面導航

摘要
著錄項
相似文獻
相關主題

摘要

Noise is inherent to low-dose CT acquisition. We propose to train a convolutional neural network (CNN) jointly with an adversarial CNN to estimate routine-dose CT images from low-dose CT images and hence reduce noise. A generator CNN was trained to transform low-dose CT images into routine-dose CT images using voxelwise loss minimization. An adversarial discriminator CNN was simultaneously trained to distinguish the output of the generator from routine-dose CT images. The performance of this discriminator was used as an adversarial loss for the generator. Experiments were performed using CT images of an anthropomorphic phantom containing calcium inserts, as well as patient non-contrast-enhanced cardiac CT images. The phantom and patients were scanned at 20% and 100% routine clinical dose. Three training strategies were compared: the first used only voxelwise loss, the second combined voxelwise loss and adversarial loss, and the third used only adversarial loss. The results showed that training with only voxelwise loss resulted in the highest peak signal-to-noise ratio with respect to reference routine-dose images. However, CNNs trained with adversarial loss captured image statistics of routine-dose images better. Noise reduction improved quantification of low-density calcified inserts in phantom CT images and allowed coronary calcium scoring in low-dose patient CT images with high noise levels. Testing took less than 10 s per CT volume. CNN-based low-dose CT noise reduction in the image domain is feasible. Training with an adversarial network improves the CNNs ability to generate images with an appearance similar to that of reference routine-dose CT images.

機譯：噪聲是低劑量CT采集所固有的。我們建議與對抗性CNN一起訓練卷積神經網絡（CNN），以從低劑量CT圖像估計常規劑量CT圖像，從而降低噪聲。訓練了CNN生成器，以使用三維像素損失最小化將低劑量CT圖像轉換為常規劑量CT圖像。同時訓練了一個對抗鑒別器CNN，以將發生器的輸出與常規劑量CT圖像區分開。該鑒別器的性能被用作發電機的對抗性損失。使用包含鈣插入物的擬人化體模的CT圖像以及患者的非增強性心臟CT圖像進行實驗。以20％和100％的常規臨床劑量對體模和患者進行掃描。比較了三種訓練策略：第一種僅使用體素丟失，第二種組合體素損失和對抗性損失，第三種僅使用對抗性損失。結果表明，相對于參考常規劑量圖像，僅以體素損失進行訓練會導致最高峰值信噪比。但是，經過對抗損失訓練的CNN可以更好地捕獲常規劑量圖像的圖像統計信息。降噪改善了幻影CT圖像中低密度鈣化插入物的定量，并允許在高噪聲水平的低劑量患者CT圖像中對冠狀動脈鈣化評分。每個CT體積的測試時間少于10秒。在圖像域中基于CNN的低劑量CT降噪是可行的。對抗網絡的訓練提高了CNN生成具有類似于參考常規劑量CT圖像外觀的圖像的能力。

著錄項

來源
《IEEE Transactions on Medical Imaging》 |2017年第12期|2536-2545|共10頁
作者
Jelmer M. Wolterink; Tim Leiner; Max A. Viergever; Ivana I?gum;
展開▼
作者單位

展開▼
收錄信息
原文格式 PDF
正文語種 eng
中圖分類
關鍵詞
Computed tomography; Generators; Noise reduction; Convolution; Training; Transforms; Calcium;

機譯：計算機斷層掃描發生器降噪卷積訓練變換鈣;
入庫時間 2022-08-17 13:03:29

相似文獻

外文文獻
中文文獻
專利

1. DN-GAN: Denoising generative adversarial networks for speckle noise reduction in optical coherence tomography images [J] . Chen Zailiang, Zeng Ziyang, Shen Hailan, Biomedical signal processing and control . 2020,第Jana期

機譯：DN-GAN：降噪生成對抗網絡，以減少光學相干斷層掃描圖像中的斑點噪聲
2. Considering anatomical prior information for low-dose CT image enhancement using attribute-augmented Wasserstein generative adversarial networks [J] . Huang Zhenxing, Liu Xinfeng, Wang Rongpin, Neurocomputing . 2021,第Mara7期

機譯：考慮使用屬性增強的Wassersein生成對沖網絡考慮了低劑量CT圖像增強的解剖學現有信息
3. Low-dose CT Image Restoration using generative adversarial networks [J] . Jongchol Kim, Jiyong Kim, Gyongwon Han, Informatics in Medicine Unlocked . 2020,第a期

機譯：低劑量CT圖像恢復使用生成的對抗網絡
4. Enhance Generative Adversarial Networks By Wavelet Transform To Denoise Low-Dose Ct Images [C] . Wanqi Su, Yili Qu, Chufu Deng, IEEE International Conference on Image Processing . 2020

機譯：通過小波變換增強生成對抗網絡以對低劑量Ct圖像進行降噪
5. Ct-Scan Image Denoising with Generative Adversarial Networks [D] . Gajera, Binit. 2020

機譯：CT掃描圖像去噪與生成對抗性網絡
6. Unpaired Low-Dose CT Denoising Network Based on Cycle-Consistent Generative Adversarial Network with Prior Image Information [O] . Chao Tang, Jie Li, Linyuan Wang, 2019

機譯：基于先驗圖像信息的周期一致生成對抗網絡的不成對小劑量CT降噪網絡
7. Residual Channel Attention Generative Adversarial Network for Image Super-Resolution and Noise Reduction [O] . Jie Cai, Zibo Meng, Chiu Man Ho 2020

機譯：用于圖像超分辨率和降噪的殘余通道注意力抗逆性網絡

站內服務

Generative Adversarial Networks for Noise Reduction in Low-Dose CT

摘要

著錄項

相似文獻

相關主題