Denoising Of Magnetic Resonance Images Research Articles

A foundation model for enhancing magnetic resonance images and downstream segmentation, registration and diagnostic tasks.

In structural magnetic resonance (MR) imaging, motion artefacts, low resolution, imaging noise and variability in acquisition protocols frequently degrade image quality and confound downstream analyses. Here we report a foundation model for the motion correction, resolution enhancement, denoising and harmonization of MR images. Specifically, we trained a tissue-classification neural network to predict tissue labels, which are then leveraged by a 'tissue-aware' enhancement network to generate high-quality MR images. We validated the model's effectiveness on a large and diverse dataset comprising 2,448 deliberately corrupted images and 10,963 images spanning a wide age range (from foetuses to elderly individuals) acquired using a variety of clinical scanners across 19 public datasets. The model consistently outperformed state-of-the-art algorithms in improving the quality of MR images, handling pathological brains with multiple sclerosis or gliomas, generating 7-T-like images from 3 T scans and harmonizing images acquired from different scanners. The high-quality, high-resolution and harmonized images generated by the model can be used to enhance the performance of models for tissue segmentation, registration, diagnosis and other downstream tasks.

FONDUE: Robust resolution-invariant denoising of MR Images using Nested UNets

Abstract Recent human magnetic resonance imaging (MRI) studies continually push the boundaries of spatial resolution as a means to enhance levels of neuroanatomical detail and increase the accuracy and sensitivity of derived brain morphometry measures. However, acquisitions required to achieve these resolutions have a higher noise floor, potentially impacting segmentation and morphometric analysis results. This study proposes a novel, fast, robust, and resolution-invariant deep learning method to denoise structural human brain MRIs. We explore denoising of T1-weighted (T1w) brain images from varying field strengths (1.5T to 7T), voxel sizes (1.2mm to 250µm), scanner vendors (Siemens, GE, and Phillips), and diseased and healthy participants from a wide age range (young adults to aging individuals). Our proposed Fast-Optimized Network for Denoising through residual Unified Ensembles (FONDUE) method demonstrated stable denoising capabilities across multiple resolutions with performance on par or superior to the state-of-the-art methods while being several orders of magnitude faster at low relative cost when using a dedicated Graphics Processing Unit (GPU). FONDUE achieved the best performance on at least one of the four denoising-performance metrics on all the test datasets used, showing its generalization capabilities and stability. Due to its high-quality performance, robustness, fast execution times, and relatively low-GPU memory requirements, as well as its open-source public availability, FONDUE can be widely used for structural MRI denoising, especially in large-cohort studies. We have made the FONDUE repository and all training and evaluation scripts as well as the trained weights available at https://github.com/waadgo/FONDUE.

An enhanced network for brain MR image denoising

Magnetic Resonance Imaging (MRI) is a cornerstone of modern medical diagnosis due to its ability to visualize intricate soft tissues without ionizing radiation. However, noise artifacts significantly degrade image quality, hindering accurate diagnosis. Traditional denoising methods struggle to preserve details while effectively reducing noise. While deep learning approaches show promise, they often focus on local information, neglecting long-range dependencies. To address these limitations, this study proposes the deep and shallow feature fusion denoising network (DAS-FFDNet) for MRI denoising. DAS-FFDNet combines shallow and deep feature extraction with a tailored fusion module, effectively capturing both local and global image information. This approach surpasses existing methods in preserving details and reducing noise, as demonstrated on publicly available T1-weighted and T2-weighted brain image datasets. The proposed model offers a valuable tool for enhancing MRI image quality and subsequent analyses.

Global and local feature extraction based on convolutional neural network residual learning for MR image denoising.

Given the different noise distribution information of global and local magnetic resonance (MR) images, this study aims to extend the current work on convolutional neural networks that preserve global structure and local details in MR image denoising tasks. This study proposed a parallel and serial network for denoising 3D MR images, called 3D-PSNet. We use the residual depthwise separable convolution block to learn the local information of the feature map, reduce the network parameters, and thus improve the training speed and parameter efficiency. In addition, we consider the feature extraction of the global image and utilize residual dilated convolution to process the feature map to expand the receptive field of the network and avoid the loss of global information. Finally, we combine both of them to form a parallel network. What's more, we integrate reinforced residual convolution blocks with dense connections to form serial network branches, which can remove redundant information and refine features to further obtain accurate noise information. The peak signal-to-noise ratio, structural similarity index measure, and root mean square error metrics of 3D-PSNet are as high as 47.79%, 99.81%, and 0.40%, respectively, achieving competitive denoising effect on three public datasets. The ablation experiments demonstrated the effectiveness of all the designed modules regarding all the evaluated metrics in both datasets. The proposed 3D-PSNet takes advantage of multi-scale receptive fields, local feature extraction and residual dense connections to more effectively restore the global structure and local fine features in MR images, and is expected to help doctors quickly and accurately diagnose patients' conditions.

Uncertainty Estimation and Out-of-Distribution Detection for Deep Learning-Based Image Reconstruction Using the Local Lipschitz.

Accurate image reconstruction is at the heart of diagnostics in medical imaging. Supervised deep learning-based approaches have been investigated for solving inverse problems including image reconstruction. However, these trained models encounter unseen data distributions that are widely shifted from training data during deployment. Therefore, it is essential to assess whether a given input falls within the training data distribution. Current uncertainty estimation approaches focus on providing an uncertainty map to radiologists, rather than assessing the training distribution fit. In this work, we propose a method based on the local Lipschitz metric to distinguish out-of-distribution images from in-distribution with an area under the curve of 99.94% for True Positive Rate versus False Positive Rate. We demonstrate a very strong relationship between the local Lipschitz value and mean absolute error (MAE), supported by a Spearman's rank correlation coefficient of 0.8475, to determine an uncertainty estimation threshold for optimal performance. Through the identification of false positives, we demonstrate the local Lipschitz and MAE relationship can guide data augmentation and reduce uncertainty. Our study was validated using the AUTOMAP architecture for sensor-to-image Magnetic Resonance Imaging (MRI) reconstruction. We demonstrate our approach outperforms baseline techniques of Monte-Carlo dropout and deep ensembles as well as the state-of-the-art Mean Variance Estimation network approach. We expand our application scope to MRI denoising and Computed Tomography sparse-to-full view reconstructions using UNET architectures. We show our approach is applicable to various architectures and applications, especially in medical imaging, where preserving diagnostic accuracy of reconstructed images remains paramount.

Wavelet-based Denoising of Magnetic Resonance Images Using Optimized Exponential Function Thresholding and Wiener Filter

Wavelet-based Denoising of Magnetic Resonance Images Using Optimized Exponential Function Thresholding and Wiener Filter

Self-supervised MRI denoising: leveraging Stein’s unbiased risk estimator and spatially resolved noise maps

Thermal noise caused by the imaged object is an intrinsic limitation in magnetic resonance imaging (MRI), resulting in an impaired clinical value of the acquisitions. Recently, deep learning (DL)-based denoising methods achieved promising results by extracting complex feature representations from large data sets. Most approaches are trained in a supervised manner by directly mapping noisy to noise-free ground-truth data and, therefore, require extensive paired data sets, which can be expensive or infeasible to obtain for medical imaging applications. In this work, a DL-based denoising approach is investigated which operates on complex-valued reconstructed magnetic resonance (MR) images without noise-free target data. An extension of Stein’s unbiased risk estimator (SURE) and spatially resolved noise maps quantifying the noise level with pixel accuracy were employed during the training process. Competitive denoising performance was achieved compared to supervised training with mean squared error (MSE) despite optimizing the model without noise-free target images. The proposed DL-based method can be applied for MR image enhancement without requiring noise-free target data for training. Integrating the noise maps as an additional input channel further enables the regulation of the desired level of denoising to adjust to the preference of the radiologist.

Denoising of magnetic resonance images of brain tumor using BT-Autonet

While obtaining medical images from sources such as Magnetic Resonance Imaging (MRI), Computed Tomography(CT), and ultrasound, noise is observed within images obtained from real world situations. Often this noise is caused due to vibrations of magnetic coils caused by quick electrical pulses, random thermal motion of protons in the tissue, reverberation and refraction artifacts. Denoising technique is one of the critical aspects in the Computer-aided Diagnosis (CAD) system, since MRI is susceptible to noises like Gaussian, Rician and Rayleigh. Traditional methods for MRI denoising are prone to challenges such as loss of information, loss caused during compression and retention of edge features. Hence, this paper presents a comparative analysis of various image denoising methods and hence, proposes an autoencoder based network Brain Tumor (BT)-Autonet for the removal of noise from brain MRI. Further, the performance analysis of the various denoising approaches is measured using different metrics. The proposed network BT-Autonet for 128 × 128 image dataset achieves a Peak Signal-to-Noise Ratio (PSNR) of 30.788, Mean Square Error(MSE) of 25.179, Structural Similarity Index Measure(SSIM) of 0.9 for Gaussian Dataset. It achieves a PSNR of 27.952, MSE of 23.129, SSIM of 0.861 for Rician Dataset and PSNR of 25.329, MSE of 44.378, SSIM of 0.873 for Rayleigh Dataset with an execution time of 10.5 s for Gaussian Dataset, 11 s for Rician Dataset and 11 s for Rayleigh Dataset. For 256 × 256 image dataset, BT-Autonet achieves a PSNR of 30.452, MSE of 30.036, SSIM of 0.816 for Gaussian Dataset while PSNR of 29.64, MSE of 41.684, SSIM of 0.809 for Rician Dataset and PSNR of 12.818, MSE of 67.219, SSIM 0.279 for Rayleigh Dataset with an execution time of 25 s for Gaussian Dataset, 27 s for Rician Dataset and 26 s for Rayleigh Dataset during the examination. Therefore, the proposed network outperformed the existing models in PSNR, SSIM, MSE and execution time.

Magnetic Resonance Image Denoising Based on Laplacian Prior Sparsity Constraint and Nonconvex Second-Order TV Penalty

Magnetic resonance (MR) imaging is considered as a very powerful imaging modality in clinical examination, but the process of image acquisition and transmission will be affected by noise, resulting in the degradation of imaging quality. In this paper, based on the Laplacian prior sparsity constraint and the nonconvex second order total variation (TV) penalty, we propose a MR images denoising model which consists of three terms. Specifically, in the first term, we use the L2-norm as the fidelity term to control the proximity between the observed image and the recovered MR image. Then, we introduce the Laplacian sparse prior constraint as the second term to mitigate the staircase artifacts in the recovered image. In the third term, we adopt the nonconvex second-order TV penalty to preserve important textures and edges. Finally, we use the alternating direction method of multipliers to solve the corresponding minimization problem. Comparative experiments on clinical data demonstrate the effectiveness of our approach in terms of PSNR and SSIM values.

Denoising of volumetric magnetic resonance imaging using multi-channel three-dimensional convolutional neural network with applications on fast spin echo acquisitions.

Three-dimensional (3D) magnetic resonance imaging (MRI) can be acquired with a high spatial resolution with flexibility being reformatted into arbitrary planes, but at the cost of reduced signal-to-noise ratio. Deep-learning methods are promising for denoising in MRI. However, the existing 3D denoising convolutional neural networks (CNNs) rely on either a multi-channel two-dimensional (2D) network or a single-channel 3D network with limited ability to extract high dimensional features. We aim to develop a deep learning approach based on multi-channel 3D convolution to utilize inherent noise information embedded in multiple number of excitation (NEX) acquisition for denoising 3D fast spin echo (FSE) MRI. A multi-channel 3D CNN is developed for denoising multi-NEX 3D FSE magnetic resonance (MR) images based on the feature extraction of 3D noise distributions embedded in 2-NEX 3D MRI. The performance of the proposed approach was compared to several state-of-the-art MRI denoising methods on both synthetic and real knee data using 2D and 3D metrics of peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM). The proposed method achieved improved denoising performance compared to the current state-of-the-art denoising methods in both slice-by-slice 2D and volumetric 3D metrics of PSNR and SSIM. A multi-channel 3D CNN is developed for denoising of multi-NEX 3D FSE MR images. The superior performance of the proposed multi-channel 3D CNN in denoising multi-NEX 3D MRI demonstrates its potential in tasks that require the extraction of high-dimensional features.

Denoising of three-dimensional fast spin echo magnetic resonance images of knee joints using spatial-variant noise-relevant residual learning of convolution neural network

PurposeTwo-dimensional (2D) fast spin echo (FSE) techniques play a central role in the clinical magnetic resonance imaging (MRI) of knee joints. Moreover, three-dimensional (3D) FSE provides high-isotropic-resolution magnetic resonance (MR) images of knee joints, but it has a reduced signal-to-noise ratio compared to 2D FSE. Deep-learning denoising methods are a promising approach for denoising MR images, but they are often trained using synthetic noise due to challenges in obtaining true noise distributions for MR images. In this study, inherent true noise information from two number of excitations (2-NEX) acquisition was used to develop a deep-learning model based on residual learning of convolutional neural network (CNN), and this model was used to suppress the noise in 3D FSE MR images of knee joints. MethodsA deep learning-based denoising method was developed. The proposed CNN used two-step residual learning over parallel transporting and residual blocks and was designed to comprehensively learn real noise features from 2-NEX training data. ResultsThe results of an ablation study validated the network design. The new method achieved improved denoising performance of 3D FSE knee MR images compared with current state-of-the-art methods, based on the peak signal-to-noise ratio and structural similarity index measure. The improved image quality after denoising using the new method was verified by radiological evaluation. ConclusionA deep CNN using the inherent spatial-varying noise information in 2-NEX acquisitions was developed. This method showed promise for clinical MRI assessments of the knee, and has potential applications for the assessment of other anatomical structures.

Denoising 3D magnetic resonance images based on weighted tensor nuclear norm minimization using balanced nonlocal patch tensors

The denoising of magnetic resonance (MR) images is important to improve the accuracy of organ tissue information recognition in the process of medical diagnosis. This paper proposes a 3D MR image denoising method based on weighted tensor nuclear norm minimization using balanced nonlocal patch tensors. In order to make utilization of the nonlocal self-similarity in high-dimensional images and the correlation among different dimensions, a high-dimensional adaptive clustering technology is developed to construct highly correlated 3D MR image patches into nonlocal patch tensors. Since the generated tensor exhibits low-rank characteristics, a low-rank tensor approximation technique based on the tensor singular value decomposition framework can be used to denoise the MR data. To improve the denoising accuracy, the local noise level information of different nonlocal patch tensors is considered. Besides, a method for constructing more balanced 3D nonlocal patch tensors is proposed to exploit the effectiveness of the tensor singular value decomposition. Experimental results show that the proposed method has a greater improvement over the existing MR image denoising methods in terms of objective measurement and subjective visual quality on simulated and real 3D MR images with different categories and noise levels.

Deep Adaptive Blending Network for 3D Magnetic Resonance Image Denoising.

The visual quality of magnetic resonance images (MRIs) is crucial for clinical diagnosis and scientific research. The main source of quality degradation is the noise generated during MRI acquisition. Although denoising MRI by deep learning methods shows great superiority compared with traditional methods, the deep learning methods reported to date in the literature cannot simultaneously leverage long-range and hierarchical information, and cannot adequately utilize the similarity in 3D MRI. In this paper, we address the two issues by proposing a deep adaptive blending network (DABN) characterized by a large receptive field residual dense block and an adaptive blending method. We first propose the large receptive field residual dense block that can capture long-range information and fuse hierarchical features simultaneously. Then we propose the adaptive blending method that produces denoised pixels by adaptively filtering 3D MRI, which explicitly utilizes the similarity in 3D MRI. Residual is also considered as a compensating item after adaptive filtering. The blending adaptive filter and residual are predicted by a network consisting of several large receptive field residual dense blocks. Experimental results show that the proposed DABN outperforms state-of-the-art denoising methods in both clinical and simulated MRI data.

Computer-aided automatic approach for denoising of magnetic resonance images

ABSTRACT This paper proposes a dual path deep convolution network based on discriminative learning for denoising MR images. The noise in MR images causes problems in identifying the regions of interest. The proposed approach is incorporated using depthwise separable convolution and local response normalisation in one path. Another path of the network is implemented using depthwise separable convolution and group normalisation. The proposed network was used to denoise the images of different noise levels, and it yields better performance as compared with various networks. The network produces clinically relevant results without retraining the network on other datasets as depicted by the evaluation metrics. The evaluation metrics improved remarkably in the results obtained by the proposed model. The additional external clinical validation was performed by the senior radiologists, and the results were found to be satisfactory for medical diagnosis. The statistical analysis of the results proves the suitability of the results for medical analysis. Further, the images were segmented for checking the applicability of the results in the biomedical domain and a remarkable improvement of about (7 ± 0.03) % and (6.5 ± 0.02) was observed in mIoU and BF score.

Denoising of magnetic resonance images using discriminative learning-based deep convolutional neural network.

The noise in magnetic resonance (MR) images causes severe issues for medical diagnosis purposes. In this paper, we propose a discriminative learning based convolutional neural network denoiser to denoise the MR image data contaminated with noise. The proposed method incorporates the use of depthwise separable convolution along with local response normalization with modified hyperparameters and internal skip connections to denoise the contaminated MR images. Moreover, the addition of parametric RELU instead of normal conventional RELU in our proposed architecture gives more stable and fine results. The denoised images were further segmented to test the appropriateness of the results. The network is trained on one dataset and tested on other dataset produces remarkably good results. Our proposed network was used to denoise the images of different noise levels, and it yields better performance as compared with various networks. The SSIM and PSNR showed an average improvement of (7.2 ± 0.002) % and (8.5 ± 0.25) % respectively when tested on different datasets without retaining the network. An improvement of 5% and 6% was achieved in the values of mean intersection over union (mIoU) and BF score when the denoised images were segmented for testing the relevancy in biomedical imaging applications. The statistical test suggests that the obtained results are statistically significant as p< 0.05. The denoised images obtained are more clinically suitable for medical image diagnosis purposes, as depicted by the evaluation parameters. Further, external clinical validation was performed by an experienced radiologist for testing the validation of the resulting images.

GPU Accelerated Bilateral Filter for MR Image Denoising

Background: Magnetic resonance (MR) imaging plays a significant role in the computer- aided diagnostic systems for remote healthcare. In such systems, the soft textures and tissues within the denoised MR image are classified by the segmentation stage using machine learning algorithms like Hidden Markov Model. Thus, the quality of the MR image is of extreme importance and is decisive in the accuracy of the process of classification and diagnosis. Objective: To provide real-time medical diagnostics in the remote healthcare intelligent setups, the research work proposes CUDA GPU based accelerated bilateral filter for fast denoising of 2D high- resolution knee MR images. Methods: To achieve optimized GPU performance with better speed-up, the work implements an improvised technique that uses on-chip shared memory in combination with a constant cache. Results: The speed-up of 382x is achieved with the new proposed optimization technique which is 2.7x as that obtained with the shared memory only approach. The superior speed-up is along with 90.6%occupancy index indicating effective parallelization. The work here also aims at justifying the appropriateness of bilateral filter over other filters for denoising magnetic resonance images. All the patents related to GPU based image denoising are revised and uniqueness of the proposed technique is confirmed. Conclusion: The results indicate that even for a 64Mpixel image, the execution time of the proposed implementation is 334.91 msec only, making the performance almost real time. This will surely contribute to the real-time computer-aided data diagnostics requirement under remote critical conditions.

Denoising of 3D Brain MR Images with Parallel Residual Learning of Convolutional Neural Network Using Global and Local Feature Extraction.

Magnetic resonance (MR) images often suffer from random noise pollution during image acquisition and transmission, which impairs disease diagnosis by doctors or automated systems. In recent years, many noise removal algorithms with impressive performances have been proposed. In this work, inspired by the idea of deep learning, we propose a denoising method named 3D-Parallel-RicianNet, which will combine global and local information to remove noise in MR images. Specifically, we introduce a powerful dilated convolution residual (DCR) module to expand the receptive field of the network and to avoid the loss of global features. Then, to extract more local information and reduce the computational complexity, we design the depthwise separable convolution residual (DSCR) module to learn the channel and position information in the image, which not only reduces parameters dramatically but also improves the local denoising performance. In addition, a parallel network is constructed by fusing the features extracted from each DCR module and DSCR module to improve the efficiency and reduce the complexity for training a denoising model. Finally, a reconstruction (REC) module aims to construct the clean image through the obtained noise deviation and the given noisy image. Due to the lack of ground-truth images in the real MR dataset, the performance of the proposed model was tested qualitatively and quantitatively on one simulated T1-weighted MR image dataset and then expanded to four real datasets. The experimental results show that the proposed 3D-Parallel-RicianNet network achieves performance superior to that of several state-of-the-art methods in terms of the peak signal-to-noise ratio, structural similarity index, and entropy metric. In particular, our method demonstrates powerful abilities in both noise suppression and structure preservation.

Boosting Magnetic Resonance Image Denoising With Generative Adversarial Networks

Denoising plays an important role in the Magnetic Resonance Imaging (MRI) applications for medical diagnosis. MRI images usually contain undesired noises which would negatively affect the exactitude of pathological diagnosis. Recently, many models for MRI denoising have been developed from deep learning networks. In this paper, we propose a novel MRI image denoising method using the conditional Generative Adversarial Networks (GANs). Specifically, a Convolutional Neural Network (CNN) is utilized as the discriminator in the process to distinguish whether the image pair obtained from the conditional GAN is a real pair which consists of a noisy image and a noise-free image or a fake pair which, on the other hand, contains a noisy image and a denoised image. In our design, the convolutional encoder-decoder networks-based generator is used to remove the noise in the noisy MRI images as much as possible. The whole architecture is trained by adversarial learning. Experiments using both synthetic and real clinical MRI datasets are conducted. When tested on the synthetic T1w images with 10% noise level, our method performed better in terms of reaching a high structural similarity index (SSIM) at 0.9489 while that of the next best method was only 0.7485. Moreover, when the image noise level was increased from 1% to 10%, our method was more stable that the SSIM only dropped about 3.2% while that of the next best method dropped about 23.7%. Simulation results demonstrate that the proposed method is more robust and outperforms the conventional methods in both the denoising level and preservation of the anatomical structures and defined contrast.

FFA-DMRI: A Network Based on Feature Fusion and Attention Mechanism for Brain MRI Denoising.

Magnetic Resonance Imaging (MRI) is an indispensable tool in the diagnosis of brain diseases due to painlessness and safety. Nevertheless, Rician noise is inevitably injected during the image acquisition process, which leads to poor observation and interferes with the treatment. Owing to the complexity of Rician noise, using the elimination method of Gaussian to remove it does not perform well. Therefore, the feature fusion and attention network (FFA-DMRI) is proposed to separate noise from observed MRI. Inspired by the attention-guided CNN network (ADNet) and Convolutional block attention module (CBAM), a spatial attention mechanism has been specially designed to obtain the area of interest in MRI. Furthermore, the feature fusion block concatenates local with global information, which makes full use of the multilevel structure and boosts the expressive ability of network. The comprehensive experiments on Alzheimer’s disease neuroimaging initiative dataset (ADNI) have demonstrated high effectiveness of FFA-DMRI with maintaining the crucial brain details. Moreover, in terms of visual inspections, the denoising results are also consistent with human perception.

CNN-DMRI: A Convolutional Neural Network for Denoising of Magnetic Resonance Images

Magnetic Resonance Images (MRI) are often contaminated by rician noise at the acquisition time. This type of noise typically deteriorates the performance of disease diagnosis by a human observer or an automated system. Thus, it is necessary to remove the rician noise from MRI scans as a preprocessing step. In this letter, we propose a novel Convolutional Neural Network (CNN), viz. CNN-DMRI, for denoising of MRI scans. The network uses a set of convolutions to separate the image features from the noise. The network also employs encoder-decoder structure for preserving the prominent features of the image while ignoring unnecessary ones. The training of the network is carried out in an end-to-end way by utilizing residual learning scheme. The performance of the proposed CNN has been tested qualitatively and quantitatively on one simulated and four real MRI datasets. Extensive experimental findings suggest that the proposed network can denoise MRI images effectively without losing crucial image details.