Hformer:highly efficient vision transformer for low-dose CT denoising

In this paper,we propose Hformer,a novel supervised learning model for low-dose computer tomography(LDCT)denoising.Hformer combines the strengths of convolutional neural networks for local feature extraction and transformer models for global feature capture.The performance of Hformer was verified and evaluated based on the AAPM-Mayo Clinic LDCT Grand Challenge Dataset.Compared with the former representative state-of-the-art(SOTA)model designs under different architectures,Hformer achieved optimal metrics without requiring a large number of learning parameters,with metrics of33.4405 PSNR,8.6956 RMSE,and 0.9163 SSIM.The experiments demonstrated designed Hformer is a SOTA model for noise suppression,structure preservation,and lesion detection.

Low-Dose CT Image Denoising Based on Improved WGAN-gp

In order to improve the quality of low-dose computational tomography (CT)images, the paper proposes an improved image denoising approach based on WGAN-gpwith Wasserstein distance. For improving the training and the convergence efficiency, thegiven method introduces the gradient penalty term to WGAN network. The novelperceptual loss is introduced to make the texture information of the low-dose imagessensitive to the diagnostician eye. The experimental results show that compared with thestate-of-art methods, the time complexity is reduced, and the visual quality of low-doseCT images is significantly improved.

Strided Self-Supervised Low-Dose CT Denoising for Lung Nodule Classification

Lung nodule classification based on low-dose computed tomography(LDCT)images has attracted major attention thanks to the reduced radiation dose and its potential for early diagnosis of lung cancer from LDCT-based lung cancer screening.However,LDCT images suffer from severe noise,largely influencing the performance of lung nodule classification.Current methods combining denoising and classification tasks typically require the corresponding normal-dose CT(NDCT)images as the supervision for the denoising task,which is impractical in the context of clinical diagnosis using LDCT.To jointly train these two tasks in a unified framework without the NDCT images,this paper introduces a novel self-supervised method,termed strided Noise2Neighbors or SN2N,for blind medical image denoising and lung nodule classification,where the supervision is generated from noisy input images.More specifically,the proposed SN2N can construct the supervision infor-mation from its neighbors for LDCT denoising,which does not need NDCT images anymore.The proposed SN2N method enables joint training of LDCT denoising and lung nodule classification tasks by using self-supervised loss for denoising and cross-entropy loss for classification.Extensively experimental results on the Mayo LDCT dataset demonstrate that our SN2N achieves competitive performance compared with the supervised learning methods that have paired NDCT images as supervision.Moreover,our results on the LIDC-IDRI dataset show that the joint training of LDCT denoising and lung nodule classification significantly improves the performance of LDCT-based lung nodule classification.

A selective kernel-based cycle-consistent generative adversarial network for unpaired low-dose CT denoising

Low-dose computed tomography(LDCT)denoising is an indispensable procedure in the medical imaging field,which not only improves image quality,but can mitigate the potential hazard to patients caused by routine doses.Despite the improvement in performance of the cycle-consistent generative adversarial network(CycleGAN)due to the well-paired CT images shortage,there is still a need to further reduce image noise while retaining detailed features.Inspired by the residual encoder–decoder convolutional neural network(RED-CNN)and U-Net,we propose a novel unsupervised model using CycleGAN for LDCT imaging,which injects a two-sided network into selective kernel networks(SK-NET)to adaptively select features,and uses the patchGAN discriminator to generate CT images with more detail maintenance,aided by added perceptual loss.Based on patch-based training,the experimental results demonstrated that the proposed SKFCycleGAN outperforms competing methods in both a clinical dataset and the Mayo dataset.The main advantages of our method lie in noise suppression and edge preservation.

Variant Wasserstein Generative Adversarial Network Applied on Low Dose CT Image Denoising

Computed Tomography(CT)images have been extensively employed in disease diagnosis and treatment,causing a huge concern over the dose of radiation to which patients are exposed.Increasing the radiation dose to get a better image may lead to the development of genetic disorders and cancer in the patients;on the other hand,decreasing it by using a Low-Dose CT(LDCT)image may cause more noise and increased artifacts,which can compromise the diagnosis.So,image reconstruction from LDCT image data is necessary to improve radiologists’judgment and confidence.This study proposed three novel models for denoising LDCT images based on Wasserstein Generative Adversarial Network(WGAN).They were incorporated with different loss functions,including Visual Geometry Group(VGG),Structural Similarity Loss(SSIM),and Structurally Sensitive Loss(SSL),to reduce noise and preserve important information on LDCT images and investigate the effect of different types of loss functions.Furthermore,experiments have been conducted on the Graphical Processing Unit(GPU)and Central Processing Unit(CPU)to compare the performance of the proposed models.The results demonstrated that images from the proposed WGAN-SSIM,WGAN-VGG-SSIM,and WGAN-VGG-SSL were denoised better than those from state-of-the-art models(WGAN,WGAN-VGG,and SMGAN)and converged to a stable equilibrium compared with WGAN and WGAN-VGG.The proposed models are effective in reducing noise,suppressing artifacts,and maintaining informative structure and texture details,especially WGAN-VGG-SSL which achieved a high peak-signalto-noise ratio(PNSR)on both GPU(26.1336)and CPU(25.8270).The average accuracy of WGAN-VGG-SSL outperformed that of the state-ofthe-art methods by 1 percent.Experiments prove that theWGAN-VGG-SSL is more stable than the other models on both GPU and CPU.

Sinogram denoising via attention residual dense convolutional neural network for low-dose computed tomography

The widespread use of computed tomography(CT)in clinical practice has made the public focus on the cumulative radiation dose delivered to patients.Low-dose CT(LDCT)reduces the X-ray radiation dose,yet compromises quality and decreases diagnostic performance.Researchers have made great efforts to develop various algorithms for LDCT and introduced deep-learning techniques,which have achieved impressive results.However,most of these methods are directly performed on reconstructed LDCT images,in which some subtle structures and details are readily lost during the reconstruction procedure,and convolutional neural network(CNN)-based methods for raw LDCT projection data are rarely reported.To address this problem,we adopted an attention residual dense CNN,referred to as AttRDN,for LDCT sinogram denoising.First,it was aided by the attention mechanism,in which the advantages of both feature fusion and global residual learning were used to extract noise from the contaminated LDCT sinograms.Then,the denoised sinogram was restored by subtracting the noise obtained from the input noisy sinogram.Finally,the CT image was reconstructed using filtered back-projection.The experimental results qualitatively and quantitatively demonstrate that the proposed AttRDN can achieve a better performance than state-of-the-art methods.Importantly,it can prevent the loss of detailed information and has the potential for clinical application.

Quantitative evaluation of deep convolutional neural network-based image denoising for low-dose computed tomography

To minimize radiation risk,dose reduction is important in the diagnostic and therapeutic applications of computed tomography(CT).However,image noise degrades image quality owing to the reduced X-ray dose and a possible unacceptably reduced diagnostic performance.Deep learning approaches with convolutional neural networks(CNNs)have been proposed for natural image denoising;however,these approaches might introduce image blurring or loss of original gradients.The aim of this study was to compare the dose-dependent properties of a CNN-based denoising method for low-dose CT with those of other noise-reduction methods on unique CT noise-simulation images.To simulate a low-dose CT image,a Poisson noise distribution was introduced to normal-dose images while convoluting the CT unit-specific modulation transfer function.An abdominal CT of 100 images obtained from a public database was adopted,and simulated dose-reduction images were created from the original dose at equal 10-step dose-reduction intervals with a final dose of 1/100.These images were denoised using the denoising network structure of CNN(DnCNN)as the general CNN model and for transfer learning.To evaluate the image quality,image similarities determined by the structural similarity index(SSIM)and peak signal-to-noise ratio(PSNR)were calculated for the denoised images.Significantly better denoising,in terms of SSIM and PSNR,was achieved by the DnCNN than by other image denoising methods,especially at the ultra-low-dose levels used to generate the 10%and 5%dose-equivalent images.Moreover,the developed CNN model can eliminate noise and maintain image sharpness at these dose levels and improve SSIM by approximately 10%from that of the original method.In contrast,under small dose-reduction conditions,this model also led to excessive smoothing of the images.In quantitative evaluations,the CNN denoising method improved the low-dose CT and prevented over-smoothing by tailoring the CNN model.

协同感知损失和注意力机制的低剂量CT去噪

由于存在特有的量子噪声,低剂量CT去噪是一项艰巨的任务。当前主流研究使用的深度学习方法存在定性和定量指标不匹配的问题,实验结果的定量指标高,但视觉效果不好。为此,提出一种感知损失和注意力机制的低剂量CT协同去噪网络。该协同机制能够在保证视觉效果的基础上明显改善现有方法定量指标低的问题。模型在网络输入端还引入8方向的边缘检测层,可提取更丰富的纹理与结构信息,进一步提升了网络效果。针对体模数据集和真实临床数据集的实验对比结果表明,该方法相比主流工作,在视觉感受和PSNR以及SSIM指标上,均有更优异表现。

Mortality outcomes of low-dose computed tomography screening for lung cancer in urban China:a decision analysis and implications for practice

Background: Mortality outcomes in trials of low-dose computed tomography(CT) screening for lung cancer are inconsistent. This study aimed to evaluate whether CT screening in urban areas of China could reduce lung cancer mortality and to investigate the factors that associate with the screening effect.Methods: A decision tree model with three scenarios(low-dose CT screening, chest X-ray screening, and no screening) was developed to compare screening results in a simulated Chinese urban cohort(100,000 smokers aged45-80 years). Data of participant characteristics were obtained from national registries and epidemiological surveys for estimating lung cancer prevalence. The selection of other tree variables such as sensitivities and specificities of low-dose CT and chest X-ray screening were based on literature research. Differences in lung cancer mortality(primary outcome), false diagnoses, and deaths due to false diagnosis were calculated. Sensitivity analyses were performed to identify the factors that associate with the screening results and to ascertain worst and optimal screening effects considering possible ranges of the variables.Results: Among the 100,000 subjects, there were 448,541, and 591 lung cancer deaths in the low-dose CT, chest X-ray, and no screening scenarios, respectively(17.2% reduction in low-dose CT screening over chest X-ray screening and 24.2% over no screening). The costs of the two screening scenarios were 9387 and 2497 false diagnoses and 7and 2 deaths due to false diagnosis among the 100,000 persons, respectively. The factors that most influenced death reduction with low-dose CT screening over no screening were lung cancer prevalence in the screened cohort, lowdose CT sensitivity, and proportion of early-stage cancers among low-dose CT detected lung cancers. Considering all possibilities, reduction in deaths(relative numbers) with low-dose CT screening in the worst and optimal cases were16(5.4%) and 288(40.2%) over no screening, respectively.Conclusions: In terms of mortality outcomes, our findings favor conducting low-dose CT screening in urban China.However, approaches to reducing false diagnoses and optimizing important screening conditions such as enrollment criteria for screening are highly needed.

Slice-wise reconstruction for low-dose cone-beam CT using a deep residual convolutional neural network

Because of the growing concern over the radiation dose delivered to patients, X-ray cone-beam CT(CBCT) imaging of low dose is of great interest. It is difficult for traditional reconstruction methods such as Feldkamp to reduce noise and keep resolution at low doses. A typical method to solve this problem is using optimizationbased methods with careful modeling of physics and additional constraints. However, it is computationally expensive and very time-consuming to reach an optimal solution. Recently, some pioneering work applying deep neural networks had some success in characterizing and removing artifacts from a low-dose data set. In this study,we incorporate imaging physics for a cone-beam CT into a residual convolutional neural network and propose a new end-to-end deep learning-based method for slice-wise reconstruction. By transferring 3D projection to a 2D problem with a noise reduction property, we can not only obtain reconstructions of high image quality, but also lower the computational complexity. The proposed network is composed of three serially connected sub-networks: a cone-to-fan transformation sub-network, a 2D analytical inversion sub-network, and an image refinement sub-network. This provides a comprehensive solution for end-to-end reconstruction for CBCT. The advantages of our method are that the network can simplify a 3D reconstruction problem to a 2D slice-wise reconstruction problem and can complete reconstruction in an end-to-end manner with the system matrix integrated into the network design. Furthermore, reconstruction can be less computationally expensive and easily parallelizable compared with iterative reconstruction methods.

用于低剂量CT图像去噪的多级双树复小波网络

基于卷积神经网络(CNN)的图像去噪方法能有效去除低剂量计算机断层扫描(CT)图像伴随的伪影和噪声,从而确保CT设备输出高质量图像同时降低辐射,这对患者健康和医学诊断具有重要意义。为了进一步提高低剂量CT图像的质量,提出一种小波域去噪网络MDTNet。首先,基于双树复小波变换(DTCWT)构造多级编解码去噪网络,在多个尺度上提取特征以保留更多高频细节;然后,利用扩展的像素重排技术替代卷积上下采样,实现多级输入和特征融合,从而降低计算复杂度;最后,通过大量训练找到最佳的去噪模型,即二级MDTNet配合LeGall滤波器和Qshift_b滤波器,并选择较大尺寸的CT图像作为训练数据。使用AAPM数据集评估MDTNet的性能,实验结果表明,MDTNet能有效去除条纹状伪影和噪声,在定量和定性评估中性能均优于同类型去噪方法。与FWDNet相比,对于1 mm的切片,MDTNet的平均峰值信噪比(PSNR)和结构相似性指数(SSIM)分别提高了0.0887 dB和0.0024;对于3 mm的切片,分别提升了0.1443 dB和0.003。对于单张512×512像素的低剂量CT图像去噪,MDTNet在GPU上仅需0.193 s。MDTNet在保持高效率的同时保留了更多的高频细节,能够为低剂量CT图像去噪提供一种新的框架。

医学CT序列图像的混合去噪算法

医学CT(computer tomography)序列图像会因为各种原因而掺杂噪声,去噪能够有效地提高图像的质量.常见的去噪算法都是针对单张图像进行,考虑到CT序列图像之间具有很高的相似性,本文提出一种基于相邻图像结构相似性的混合去噪算法.该算法首先计算序列图像的最大和最小灰度值,根据灰度值绘制直方图,设定相关阈值参数,根据筛选之后的直方图计算窗宽窗位,然后进行调窗处理.之后计算目标图像与其前后相邻图像之间的结构相似性,最后根据结构相似性对3张图片混合使用BM3D和高斯滤波2种去噪算法.通过对比实验表明,该算法在均方误差、峰值信噪比和结构相似性三方面都有所提高,能够有效地提高图像质量.

基于特征融合的低剂量CT图像降噪方法

近年来低剂量CT(Low Dose CT,LDCT)被广泛应用于临床诊断中,但LDCT会产生不规则的噪声。已有的降噪方法往往缺乏对全局特征信息的考虑,以及不注重边缘特征信息和重建图像的视觉效果。为此,提出了一种基于特征融合的低剂量CT图像降噪方法。首先,利用Transformer优异的全局感受野提取图像的全局特征信息,并利用卷积神经网络(Convolutional Neural Network,CNN)良好的局部特征提取能力提取图像的局部特征信息。在Transformer模块中加入维度变换思想,以更好地抑制噪声;在CNN模块中使用稠密连接的方式将浅层网络的特征信息复用于深层网络中,以此保存更多的特征信息。其次,为了获取更加丰富的图像细节特征,使用了改进的索伯边缘增强算子来加强模型对边缘特征信息的提取能力。最后,将Transformer模块和CNN模块获取的特征信息进行融合并输出重建图像。此外,为了使降噪重建后的图像有更好的质量和视觉效果,设计了一个多尺度复合损失函数。实验表明:在AAPM-Mayo数据集的降噪实验中,与当前主流的LDCT图像降噪方法相比,本文方法取得了更好的降噪效果。

基于电火花震源的地震波CT波速影像图去噪

在南京某项目深部采空区采用电火花做为震源进行地震波CT探测时,发现波速影像图上出现诸多假异常,解译成果与实际地质情况不符。鉴于存在的问题,对地震波CT波速图像开展去噪研究,探索到一种使解译的速度值更加接近于实际值一种方法。实践表明,此技术可有效去除地震波CT图像波速影像图上的假异常,提高探测精度。研究成果可为物探专业的地震波CT技术发展提供借鉴。

Robust restoration of low-dose cerebral perfusion CT images using NCS-Unet

Cerebral perfusion computed tomography(PCT)is an important imaging modality for evaluating cerebrovascular diseases and stroke symptoms.With widespread public concern about the potential cancer risks and health hazards associated with cumulative radiation exposure in PCT imaging,considerable research has been conducted to reduce the radiation dose in X-ray-based brain perfusion imaging.Reducing the dose of X-rays causes severe noise and artifacts in PCT images.To solve this problem,we propose a deep learning method called NCS-Unet.The exceptional characteristics of non-subsampled contourlet transform(NSCT)and the Sobel filter are introduced into NCS-Unet.NSCT decomposes the convolved features into high-and low-frequency components.The decomposed high-frequency component retains image edges,contrast imaging traces,and noise,whereas the low-frequency component retains the main image information.The Sobel filter extracts the contours of the original image and the imaging traces caused by the contrast agent decay.The extracted information is added to NCS-Unet to improve its performance in noise reduction and artifact removal.Qualitative and quantitative analyses demonstrated that the proposed NCS-Unet can improve the quality of low-dose cone-beam CT perfusion reconstruction images and the accuracy of perfusion parameter calculations.

基于多尺度动态卷积和边缘增强的低剂量CT去噪

计算机断层扫描(CT)技术广泛应用于疾病检测与筛查,然而在扫描过程中产生的X射线辐射会对人体造成伤害.采用低剂量CT可以减少患者的辐射暴露,但是重建的图像会有显著的噪声和伪影,干扰医生的诊断.针对这一挑战,众多学者提出了基于传统卷积神经网络的低剂量CT去噪算法,并已取得显著成就.然而,传统卷积在不同像素位置共用相同卷积滤波器,这会忽略不同图像区域的内容差异,导致去噪结果的过度平滑化.为避免这一问题,本文提出一种基于多尺度动态卷积和边缘增强的低剂量CT去噪网络MDCEENet,旨在在去噪过程中保留更多的图像纹理和结构细节.MDCEENet是自编码器结构,包含编码器和解码器两个主要模块.具体而言,将低剂量CT图像及其边缘信息输入到编码器中,通过多尺度特征流MFS和边缘信息流EIS,分别提取多尺度图像特征和图像边缘特征,并将它们融合成引导信息GI,引导解码器中多尺度动态卷积块MDConvBlock的参数生成.在GI的引导下,MDConvBlock模块对上采样特征进行多尺度空洞卷积计算,旨在获取更高质量的重建图像.本文在Mayo Clinic公开的两个数据集上执行了相关实验,通过实验结果可知,MDCEENet的去噪效果优于DnCNN、RED-CNN、WGAN、CNCL、NBNet,获得了最优的平均峰值信噪比和平均结构相似性指标,这表明本文提出方法的优越性.本文还在这两个数据集上进行了消融实验,来说明MDCEENet中引入多尺度动态卷积和边缘信息的有效性,以及与ADFNet网络的区别.实验结果表明了本文提出的方法相比于ADFNet更适用于低剂量CT去噪任务.

低剂量CT图像降噪的深度图像先验的目标偏移加速算法

低剂量CT(LDCT)图像可大幅降低X射线辐射剂量,但存在大量噪声影响医生诊断。深度图像先验(DIP)是用随机张量作为神经网络的输入图像,以单张LDCT图像为目标进行迭代的无监督深度学习算法。但DIP方法需经过上千次的网络迭代才能得到最佳降噪结果,导致该方法运行速度过慢。因此,该文提出一种用于LDCT降噪的目标偏移DIP加速算法,旨在保持降噪图像质量的基础上提高运行速度。根据一个器官(如肺部)LDCT切片序列图像的相似性,该算法将以各切片分别作为目标图像对应的相互独立的网络迭代通过继承参数关联起来,在上一切片对应的网络参数的基础上更新当前切片对应的网络参数,并将当前切片对应的网络参数作为下一切片对应的网络迭代的基础;由于DIP网络的输入是固定的随机张量,与目标图像差距较大,该文利用传统降噪模型预处理后的LDCT图像作为网络输入,进一步提高网络迭代速度。实验表明,不使用传统模型预处理时,与原DIP网络运行速度相比,该文所提出的加速算法可以将迭代速度提高10.45%;当使用经过相对全变分(RTV)模型预处理的LDCT作为网络输入时,图像峰值信噪比不仅可以达到29.13,而且总迭代速度可以提高94.31%。综上所述,该文算法可在保持DIP降噪效果的基础上,大幅度提高运行速度,特别是RTV模型预处理后的CT图像作为网络输入时,对提高运行速度的效果更加明显。

基于深度学习的低剂量CT图像去噪方法综述

由于低剂量CT情境下医学图像存在多样的噪声,其强度和种类各异,因此选择合适的算法对去噪至关重要。传统图像去噪方法基于先验知识,其优化过程相对繁琐,存在保留图像细节和处理效率方面的一定限制。相较之下,基于深度学习的去噪方法具备学习能力强大、非线性建模、端到端学习、适应性强和大规模并行计算等独特优势,使其相对于传统方法在处理复杂噪声场景时更为有效。本文全面概括并深入分析了当前低剂量CT图像去噪方法的研究热点。首先,简要介绍了低剂量CT图像去噪的步骤和过程。其次,结合当前基于深度学习的低剂量CT图像去噪方法的研究现状,重点探讨了残差学习、注意力网络以及自监督学习这三个最具代表性的研究热点,详细阐述了各种基础网络架构及其改进方法在低剂量CT图像去噪中的应用情况。最后,总结了当前低剂量CT图像降噪方法所面临的主要挑战,并提出了未来的研究方向,以促进低剂量CT图像去噪技术的进一步发展。

基于跨尺度边缘增强深度卷积神经网络的低剂量CT图像去噪

为了提高去噪网络的可解释性,将传统滤波算子的优势融入到网络设计中,提出了基于跨尺度边缘增强的深度卷积神经网络(CEDCNN).将传统边缘算子与卷积相结合,设计出轻量化的边缘增强模块,增强边缘信息对网络结果的影响.基于自适应一致性先验算法构建深度迭代网络,进一步提取边缘增强特征,从而实现端到端可训练、可解释的深度去噪网络.将均方误差和多尺度注意残差感知损失相结合,解决了重建图像过平滑的问题.实验结果表明,CEDCNN去噪网络的PSNR、RMSE、SSIM评价指标分别为43.647 5 dB、0.006 8、0.987 5,说明该方法可显著提高去噪后的图像质量,有效保证低剂量X射线下CT图像的成像质量和精度,去噪效果与正常剂量CT图像所展现的人体组织细节相当.

基于RN-CNN模型的低剂量CT图像去噪方法

低剂量CT在降低辐射剂量、减少对人体伤害的同时CT的成像质量也显著下降。因此提出一种基于ResNet的卷积神经网络(RN-CNN),用来去除低剂量CT中所含的量子噪声。在正常剂量的CT图像上加入泊松噪声模拟低剂量CT图像,将模拟的低剂量CT输入RN-CNN对图像进行特征提取,引入残差网络与尺度不变的空间金字塔池化(S-SPP),避免梯度消失问题并增加网络有效特征;使用扩张卷积增大网络的感受野,保留图像内部数据结构,得到更好的分割效果;将低剂量CT图像和噪声图像分离得到正常剂量CT图。通过实验表明,该算法不仅能有效去除噪声,而且纹理细节也得以保留,与现有的算法相比较,RN-CNN去噪方法在PSNR指标上平均提高1.80,在SSIM指标上平均提高了0.078 3。