Collaborative Filtering of Correlated Noise: Exact Transform-Domain Variance for Improved Shrinkage and Patch Matching

Adversarial EM for variational deep learning: Application to semi-supervised image quality enhancement in low-dose PET and low-dose CT

In positron emission tomography (PET) and X-ray computed tomography (CT), reducing radiation dose can cause significant degradation in image quality. For image quality enhancement in low-dose PET and CT, we propose a novel theoretical adversarial and variational deep neural network (DNN) framework relying on expectation maximization (EM) based learning, termed adversarial EM (AdvEM). AdvEM proposes an encoder-decoder architecture with a multiscale latent space, and generalized-Gaussian models enabling datum-specific robust statistical modeling in latent space and image space. The model robustness is further enhanced by including adversarial learning in the training protocol. Unlike typical variational-DNN learning, AdvEM proposes latent-space sampling from the posterior distribution, and uses a Metropolis-Hastings scheme. Unlike existing schemes for PET or CT image enhancement which train using pairs of low-dose images with their corresponding normal-dose versions, we propose a semi-supervised AdvEM (ssAdvEM) framework that enables learning using a small number of normal-dose images. AdvEM and ssAdvEM enable per-pixel uncertainty estimates for their outputs. Empirical analyses on real-world PET and CT data involving many baselines, out-of-distribution data, and ablation studies show the benefits of the proposed framework.

Three-dimensional atomic insights into the metal-oxide interface in Zr-ZrO2 nanoparticles

Metal-oxide interfaces with poor coherency have specific properties comparing to bulk materials and offer broad applications in heterogeneous catalysis, battery, and electronics. However, current understanding of the three-dimensional (3D) atomic metal-oxide interfaces remains limited because of their inherent structural complexity and the limitations of conventional two-dimensional imaging techniques. Here, we determine the 3D atomic structure of metal-oxide interfaces in zirconium-zirconia nanoparticles using atomic-resolution electron tomography. We quantitatively analyze the atomic concentration and the degree of oxidation, and find the coherency and translational symmetry of the interfaces are broken. Atoms at the interface have low structural ordering, low coordination, and elongated bond length. Moreover, we observe porous structures such as Zr vacancies and nano-pores, and investigate their distribution. Our findings provide a clear 3D atomic picture of metal-oxide interface with direct experimental evidence. We anticipate this work could encourage future studies on fundamental problems of oxides, such as interfacial structures in semiconductor and atomic motion during oxidation process.

Utility of block-matching and 3D filter for reproducibility of lung density and denoising in low-dose chest CT: A pilot study

PURPOSE

This study aimed to acquire an image quality consistent with that of full-dose chest computed tomography (CT) when obtaining low-dose chest CT images and to analyze the effects of block-matching and 3D (BM3D) filters on lung density measurements and noise reduction in lung parenchyma.

METHODS

Using full-dose chest CT images, we evaluated lung density measurements and noise reduction in lung parenchyma images for low-dose chest CT. Three filters (median, Wiener, and the proposed BM3D) were applied to low-dose chest CT images for comparison and analysis with images from full-dose chest CT. To evaluate lung density measurements, we measured CT attenuation at the 15th percentile of the lung CT histogram. The coefficient of variation (COV) and contrast-to-noise ratio (CNR) were used to evaluate the noise level.

RESULTS

The 15th percentile of the lung CT histogram showed the smallest difference between full- and low-dose CT when applying the BM3D filter, and the highest difference between full- and low-dose CT without filters (full-dose =  - 926.28 ± 0.32, BM3D =  - 926.65 ± 0.32, and low-dose =  - 959.43 ± 0.95) (p < 0.05). The COV was smallest when applying the BM3D filter, whereas the CNR was the highest (p < 0.05).

CONCLUSIONS

The results of the study prove that the BM3D filter can reduce image noise while increasing the reproducibility of the lung density, even for low-dose chest CT.

Pure Vision Transformer (CT-ViT) with Noise2Neighbors Interpolation for Low-Dose CT Image Denoising

Convolutional neural networks (CNN) have been used for a wide variety of deep learning applications, especially in computer vision. For medical image processing, researchers have identified certain challenges associated with CNNs. These challenges encompass the generation of less informative features, limitations in capturing both high and low-frequency information within feature maps, and the computational cost incurred when enhancing receptive fields by deepening the network. Transformers have emerged as an approach aiming to address and overcome these specific limitations of CNNs in the context of medical image analysis. Preservation of all spatial details of medical images is necessary to ensure accurate patient diagnosis. Hence, this research introduced the use of a pure Vision Transformer (ViT) for a denoising artificial neural network for medical image processing specifically for low-dose computed tomography (LDCT) image denoising. The proposed model follows a U-Net framework that contains ViT modules with the integration of Noise2Neighbor (N2N) interpolation operation. Five different datasets containing LDCT and normal-dose CT (NDCT) image pairs were used to carry out this experiment. To test the efficacy of the proposed model, this experiment includes comparisons between the quantitative and visual results among CNN-based (BM3D, RED-CNN, DRL-E-MP), hybrid CNN-ViT-based (TED-Net), and the proposed pure ViT-based denoising model. The findings of this study showed that there is about 15-20% increase in SSIM and PSNR when using self-attention transformers than using the typical pure CNN. Visual results also showed improvements especially when it comes to showing fine structural details of CT images.

In-situ heating-and-electron tomography for materials research: from 3D (in-situ 2D) to 4D (in-situ 3D)

In-situ observation has expanded the application of transmission electron microscopy (TEM) and has made a significant contribution to materials research and development for energy, biomedical, quantum, etc. Recent technological developments related to in-situ TEM have empowered the incorporation of three-dimensional observation, which was previously considered incompatible. In this review article, we take up heating as the most commonly used external stimulus for in-situ TEM observation and overview recent in-situ TEM studies. Then, we focus on the electron tomography (ET) and in-situ heating combined observation by introducing the authors' recent research as an example. Assuming that in-situ heating observation is expanded from two dimensions to three dimensions using a conventional TEM apparatus and a commercially available in-situ heating specimen holder, the following in-situ heating-and-ET observation procedure is proposed: (i) use a rapid heating-and-cooling function of a micro-electro-mechanical system holder; (ii) heat and cool the specimen intermittently and (iii) acquire a tilt-series dataset when the specimen heating is stopped. This procedure is not too technically challenging and can have a wide range of applications. Essential technical points for a successful 4D (space and time) observation will be discussed through reviewing the authors' example application.

Systematic Review on Learning-based Spectral CT

Spectral computed tomography (CT) has recently emerged as an advanced version of medical CT and significantly improves conventional (single-energy) CT. Spectral CT has two main forms: dual-energy computed tomography (DECT) and photon-counting computed tomography (PCCT), which offer image improvement, material decomposition, and feature quantification relative to conventional CT. However, the inherent challenges of spectral CT, evidenced by data and image artifacts, remain a bottleneck for clinical applications. To address these problems, machine learning techniques have been widely applied to spectral CT. In this review, we present the state-of-the-art data-driven techniques for spectral CT.

Denoising and uncertainty estimation in parameter mapping with approximate Bayesian deep image priors

PURPOSE

To mitigate the problem of noisy parameter maps with high uncertainties by casting parameter mapping as a denoising task based on Deep Image Priors.

METHODS

We extend the concept of denoising with Deep Image Prior (DIP) into parameter mapping by treating the output of an image-generating network as a parametrization of tissue parameter maps. The method implicitly denoises the parameter mapping process by filtering low-level image features with an untrained convolutional neural network (CNN). Our implementation includes uncertainty estimation from Bernoulli approximate variational inference, implemented with MC dropout, which provides model uncertainty in each voxel of the denoised parameter maps. The method is modular, so the specifics of different applications (e.g., T1 mapping) separate into application-specific signal equation blocks. We evaluate the method on variable flip angle T1 mapping, multi-echo T2 mapping, and apparent diffusion coefficient mapping.

RESULTS

We found that deep image prior adapts successfully to several applications in parameter mapping. In all evaluations, the method produces noise-reduced parameter maps with decreased uncertainty compared to conventional methods. The downsides of the proposed method are the long computational time and the introduction of some bias from the denoising prior.

CONCLUSION

DIP successfully denoise the parameter mapping process and applies to several applications with limited hyperparameter tuning. Further, it is easy to implement since DIP methods do not use network training data. Although time-consuming, uncertainty information from MC dropout makes the method more robust and provides useful information when properly calibrated.

Deep learning approach for denoising low-SNR correlation plenoptic images

Correlation Plenoptic Imaging (CPI) is a novel volumetric imaging technique that uses two sensors and the spatio-temporal correlations of light to detect both the spatial distribution and the direction of light. This novel approach to plenoptic imaging enables refocusing and 3D imaging with significant enhancement of both resolution and depth of field. However, CPI is generally slower than conventional approaches due to the need to acquire sufficient statistics for measuring correlations with an acceptable signal-to-noise ratio (SNR). We address this issue by implementing a Deep Learning application to improve image quality with undersampled frame statistics. We employ a set of experimental images reconstructed by a standard CPI architecture, at three different sampling ratios, and use it to feed a CNN model pre-trained through the transfer learning paradigm U-Net architecture with VGG-19 net for the encoding part. We find that our model reaches a Structural Similarity (SSIM) index value close to 1 both for the test sample (SSIM = [Formula: see text]) and in 5-fold cross validation (SSIM = [Formula: see text]); the results are also shown to outperform classic denoising methods, in particular for images with lower SNR. The proposed work represents the first application of Artificial Intelligence in the field of CPI and demonstrates its high potential: speeding-up the acquisition by a factor 20 over the fastest CPI so far demonstrated, enabling recording potentially 200 volumetric images per second. The presented results open the way to scanning-free real-time volumetric imaging at video rate, which is expected to achieve a substantial influence in various applications scenarios, from monitoring neuronal activity to machine vision and security.

Deep Learning-Based Denoising of CEST MR Data: A Feasibility Study on Applying Synthetic Phantoms in Medical Imaging

Chemical Exchange Saturation Transfer (CEST) magnetic resonance imaging (MRI) provides a novel method for analyzing biomolecule concentrations in tissues without exogenous contrast agents. Despite its potential, achieving a high signal-to-noise ratio (SNR) is imperative for detecting small CEST effects. Traditional metrics such as Magnetization Transfer Ratio Asymmetry (MTRasym) and Lorentzian analyses are vulnerable to image noise, hampering their precision in quantitative concentration estimations. Recent noise-reduction algorithms like principal component analysis (PCA), nonlocal mean filtering (NLM), and block matching combined with 3D filtering (BM3D) have shown promise, as there is a burgeoning interest in the utilization of neural networks (NNs), particularly autoencoders, for imaging denoising. This study uses the Bloch-McConnell equations, which allow for the synthetic generation of CEST images and explores NNs efficacy in denoising these images. Using synthetically generated phantoms, autoencoders were created, and their performance was compared with traditional denoising methods using various datasets. The results underscored the superior performance of NNs, notably the ResUNet architectures, in noise identification and abatement compared to analytical approaches across a wide noise gamut. This superiority was particularly pronounced at elevated noise intensities in the in vitro data. Notably, the neural architectures significantly improved the PSNR values, achieving up to 35.0, while some traditional methods struggled, especially in low-noise reduction scenarios. However, the application to the in vivo data presented challenges due to varying noise profiles. This study accentuates the potential of NNs as robust denoising tools, but their translation to clinical settings warrants further investigation.

Improving MR image quality with a multi-task model, using convolutional losses

PURPOSE

During the acquisition of MRI data, patient-, sequence-, or hardware-related factors can introduce artefacts that degrade image quality. Four of the most significant tasks for improving MRI image quality have been bias field correction, super-resolution, motion-, and noise correction. Machine learning has achieved outstanding results in improving MR image quality for these tasks individually, yet multi-task methods are rarely explored.

METHODS

In this study, we developed a model to simultaneously correct for all four aforementioned artefacts using multi-task learning. Two different datasets were collected, one consisting of brain scans while the other pelvic scans, which were used to train separate models, implementing their corresponding artefact augmentations. Additionally, we explored a novel loss function that does not only aim to reconstruct the individual pixel values, but also the image gradients, to produce sharper, more realistic results. The difference between the evaluated methods was tested for significance using a Friedman test of equivalence followed by a Nemenyi post-hoc test.

RESULTS

Our proposed model generally outperformed other commonly-used correction methods for individual artefacts, consistently achieving equal or superior results in at least one of the evaluation metrics. For images with multiple simultaneous artefacts, we show that the performance of using a combination of models, trained to correct individual artefacts depends heavily on the order that they were applied. This is not an issue for our proposed multi-task model. The model trained using our novel convolutional loss function always outperformed the model trained with a mean squared error loss, when evaluated using Visual Information Fidelity, a quality metric connected to perceptual quality.

CONCLUSION

We trained two models for multi-task MRI artefact correction of brain, and pelvic scans. We used a novel loss function that significantly improves the image quality of the outputs over using mean squared error. The approach performs well on real world data, and it provides insight into which artefacts it detects and corrects for. Our proposed model and source code were made publicly available.

Efficient compressed sensing reconstruction for 3D fluorescence microscopy using OptoMechanical Modulation Tomography (OMMT) with a 1+2D regularization

OptoMechanical Modulation Tomography (OMMT) exploits compressed sensing to reconstruct high resolution microscopy volumes from fewer measurement images compared to exhaustive section sampling in conventional light sheet microscopy. Nevertheless, the volumetric reconstruction process is computationally expensive, making it impractically slow to use on large-size images, and prone to generating visual artefacts. Here, we propose a reconstruction approach that uses a 1+2D Total Variation (TV1+2) regularization that does not generate such artefacts and is amenable to efficient implementation using parallel computing. We evaluate our method for accuracy and scaleability on simulated and experimental data. Using a high quality, but computationally expensive, Plug-and-Play (PnP) method that uses the BM4D denoiser as a benchmark, we observe that our approach offers an advantageous trade-off between speed and accuracy.

Neuromorphic applications in medicine

In recent years, there has been a growing demand for miniaturization, low power consumption, quick treatments, and non-invasive clinical strategies in the healthcare industry. To meet these demands, healthcare professionals are seeking new technological paradigms that can improve diagnostic accuracy while ensuring patient compliance. Neuromorphic engineering, which uses neural models in hardware and software to replicate brain-like behaviors, can help usher in a new era of medicine by delivering low power, low latency, small footprint, and high bandwidth solutions. This paper provides an overview of recent neuromorphic advancements in medicine, including medical imaging and cancer diagnosis, processing of biosignals for diagnosis, and biomedical interfaces, such as motor, cognitive, and perception prostheses. For each section, we provide examples of how brain-inspired models can successfully compete with conventional artificial intelligence algorithms, demonstrating the potential of neuromorphic engineering to meet demands and improve patient outcomes. Lastly, we discuss current struggles in fitting neuromorphic hardware with non-neuromorphic technologies and propose potential solutions for future bottlenecks in hardware compatibility.

Large field-of-view thermal imaging via all-silicon meta-optics

A broad range of imaging and sensing technologies in the infrared require large field-of-view (FoV) operation. To achieve this, traditional refractive systems often employ multiple elements to compensate for aberrations, which leads to excess size, weight, and cost. For many applications, including night vision eye-wear, air-borne surveillance, and autonomous navigation for unmanned aerial vehicles, size and weight are highly constrained. Sub-wavelength diffractive optics, also known as meta-optics, can dramatically reduce the size, weight, and cost of these imaging systems, as meta-optics are significantly thinner and lighter than traditional refractive lenses. Here, we demonstrate 80° FoV thermal imaging in the long-wavelength infrared regime (8-12 µm) using an all-silicon meta-optic with an entrance aperture and lens focal length of 1 cm.

Three-dimensional nanoimaging of fuel cell catalyst layers

Catalyst layers in proton exchange membrane fuel cells consist of platinum-group-metal nanocatalysts supported on carbon aggregates, forming a porous structure through which an ionomer network percolates. The local structural character of these heterogeneous assemblies is directly linked to the mass-transport resistances and subsequent cell performance losses; its three-dimensional visualization is therefore of interest. Herein we implement deep-learning-aided cryogenic transmission electron tomography for image restoration, and we quantitatively investigate the full morphology of various catalyst layers at the local-reaction-site scale. The analysis enables computation of metrics such as the ionomer morphology, coverage and homogeneity, location of platinum on the carbon supports, and platinum accessibility to the ionomer network, with the results directly compared and validated with experimental measurements. We expect that our findings and methodology for evaluating catalyst layer architectures will contribute towards linking the morphology to transport properties and overall fuel cell performance.

DEEP LEARNING PREDICTION OF MOTOR PERFORMANCE IN STROKE INDIVIDUALS USING NEUROIMAGING DATA

The degree of motor impairment and profile of recovery after stroke are difficult to predict for each individual. Measures obtained from clinical assessments, as well as neurophysiological and neuroimaging techniques have been used as potential biomarkers of motor recovery, with limited accuracy up to date. To address this, the present study aimed to develop a deep learning model based on structural brain images obtained from stroke participants and healthy volunteers. The following inputs were used in a multi-channel 3D convolutional neural network (CNN) model: fractional anisotropy, mean diffusivity, radial diffusivity, and axial diffusivity maps obtained from Diffusion Tensor Imaging (DTI) images, white and gray matter intensity values obtained from Magnetic Resonance Imaging, as well as demographic data (e.g., age, gender). Upper limb motor function was classified into "Poor" and "Good" categories. To assess the performance of the DL model, we compared it to more standard machine learning (ML) classifiers including k-nearest neighbor, support vector machines (SVM), Decision Trees, Random Forests, Ada Boosting, and Naïve Bayes, whereby the inputs of these classifiers were the features taken from the fully connected layer of the CNN model. The highest accuracy and area under the curve values were 0.92 and 0.92 for the 3D-CNN and 0.91 and 0.91 for the SVM, respectively. The multi-channel 3D-CNN with residual blocks and SVM supported by DL was more accurate than traditional ML methods to classify upper limb motor impairment in the stroke population. These results suggest that combining volumetric DTI maps and measures of white and gray matter integrity can improve the prediction of the degree of motor impairment after stroke. Identifying the potential of recovery early on after a stroke could promote the allocation of resources to optimize the functional independence of these individuals and their quality of life.

Deep learning-based noise filtering toward millisecond order imaging by using scanning transmission electron microscopy

Application of scanning transmission electron microscopy (STEM) to in situ observation will be essential in the current and emerging data-driven materials science by taking STEM's high affinity with various analytical options into account. As is well known, STEM's image acquisition time needs to be further shortened to capture a targeted phenomenon in real-time as STEM's current temporal resolution is far below the conventional TEM's. However, rapid image acquisition in the millisecond per frame or faster generally causes image distortion, poor electron signals, and unidirectional blurring, which are obstacles for realizing video-rate STEM observation. Here we show an image correction framework integrating deep learning (DL)-based denoising and image distortion correction schemes optimized for STEM rapid image acquisition. By comparing a series of distortion corrected rapid scan images with corresponding regular scan speed images, the trained DL network is shown to remove not only the statistical noise but also the unidirectional blurring. This result demonstrates that rapid as well as high-quality image acquisition by STEM without hardware modification can be established by the DL. The DL-based noise filter could be applied to in-situ observation, such as dislocation activities under external stimuli, with high spatio-temporal resolution.

Dilated Convolution ResNet with Boosting Attention Modules and Combined Loss Functions for LDCT Image Denoising

With the increasing concern regarding the radiation exposure of patients undergoing computed tomography (CT) scans, researchers have been using deep learning techniques to improve the quality of denoised low-dose CT (LDCT) images. In this paper, a cascaded dilated residual network (ResNet) with integrated attention modules, specifically spatial- and channel- attention modules, is proposed. This experiment demonstrated how these attention modules improved the denoised CT image by testing a simple ResNet with and without the modules. Further, an investigation regarding the effectiveness of per-pixel loss, perceptual loss via VGG16-Net, and structural dissimilarity loss functions is also covered through an ablation experiment. By knowing how these loss functions affect the output denoised images, a combination of the these loss function is then proposed which aims to prevent edge over-smoothing, enhance textural details and finally, preserve structural details on the denoised images. Finally, a bench testing was also done by comparing the visual and quantitative results of the proposed model with the state-of-the-art models such as block matching 3D (BM3D), patch-GAN and dilated convolution with edge detection layer (DRL-E-MP) for accuracy.

Macular Vascular Imaging and Connectivity Analysis Using High-Resolution Optical Coherence Tomography

Purpose

To characterize macular blood flow connectivity in vivo using high-resolution optical coherence tomography (HighRes OCT).

Methods

Cross-sectional, observational study. Dense (6-µm interscan distance) perifoveal HighRes OCT raster scans were performed on healthy participants. To mitigate the limitations of projection-resolved OCT-angiography, flow and structural data were used to observe the vascular structures of the superficial vascular complex (SVC) and the deep vascular complex. Vascular segmentation and rendering were performed using Imaris 9.5 software. Inflow and outflow patterns were classified according to vascular diameter and branching order from superficial arteries and veins, respectively.

Results

Eight eyes from eight participants were included in this analysis, from which 422 inflow and 459 outflow connections were characterized. Arteries had direct arteriolar connections to the SVC (78%) and to the intermediate capillary plexus (ICP, 22%). Deep capillary plexus (DCP) inflow derived from small-diameter vessels succeeding ICP arterioles. The most prevalent outflow pathways coursed through superficial draining venules (74%). DCP draining venules ordinarily merged with ICP draining venules and drained independently of superficial venules in 21% of cases. The morphology of DCP draining venules in structural HighRes OCT is distinct from other vessels crossing the inner nuclear layer and can be used to identify superficial veins.

Conclusions

Vascular connectivity analysis supports a hybrid circuitry of blood flow within the human parafoveal macula.

Translational Relevance

Characterization of parafoveal macular blood flow connectivity in vivo using a precise segmentation of HighRes OCT is consistent with ground-truth microscopy studies and shows a hybrid circuitry.

Framework for denoising Monte Carlo photon transport simulations using deep learning

SIGNIFICANCE

The Monte Carlo (MC) method is widely used as the gold-standard for modeling light propagation inside turbid media, such as human tissues, but combating its inherent stochastic noise requires one to simulate a large number photons, resulting in high computational burdens.

AIM

We aim to develop an effective image denoising technique using deep learning (DL) to dramatically improve the low-photon MC simulation result quality, equivalently bringing further acceleration to the MC method.

APPROACH

We developed a cascade-network combining DnCNN with UNet, while extending a range of established image denoising neural-network architectures, including DnCNN, UNet, DRUNet, and deep residual-learning for denoising MC renderings (ResMCNet), in handling three-dimensional MC data and compared their performances against model-based denoising algorithms. We also developed a simple yet effective approach to creating synthetic datasets that can be used to train DL-based MC denoisers.

RESULTS

Overall, DL-based image denoising algorithms exhibit significantly higher image quality improvements over traditional model-based denoising algorithms. Among the tested DL denoisers, our cascade network yields a 14 to 19 dB improvement in signal-to-noise ratio, which is equivalent to simulating 25  ×   to 78  ×   more photons. Other DL-based methods yielded similar results, with our method performing noticeably better with low-photon inputs and ResMCNet along with DRUNet performing better with high-photon inputs. Our cascade network achieved the highest quality when denoising complex domains, including brain and mouse atlases.

CONCLUSIONS

Incorporating state-of-the-art DL denoising techniques can equivalently reduce the computation time of MC simulations by one to two orders of magnitude. Our open-source MC denoising codes and data can be freely accessed at http://mcx.space/.

Ring artifact and Poisson noise attenuation via volumetric multiscale nonlocal collaborative filtering of spatially correlated noise

X-ray micro-tomography systems often suffer from high levels of noise. In particular, severe ring artifacts are common in reconstructed images, caused by defects in the detector, calibration errors, and fluctuations producing streak noise in the raw sinogram data. Furthermore, the projections commonly contain high levels of Poissonian noise arising from the photon-counting detector. This work presents a 3-D multiscale framework for streak attenuation through a purposely designed collaborative filtering of correlated noise in volumetric data. A distinct multiscale denoising step for attenuation of the Poissonian noise is further proposed. By utilizing the volumetric structure of the projection data, the proposed fully automatic procedure offers improved feature preservation compared with 2-D denoising and avoids artifacts which arise from individual filtering of sinograms.

Noise-immune image blur detection via sequency spectrum truncation

Blur detection is aimed to differentiate the blurry and sharp regions from a given image. This task has attracted much attention in recent years due to its importance in computer vision with the integration of image processing and artificial intelligence. However, blur detection still suffers from problems such as the oversensitivity to image noise and the difficulty in cost–benefit balance. To deal with these issues, we propose an accurate and efficient blur detection method, which is concise in architecture and robust against noise. First, we develop a sequency spectrum-based blur metric to estimate the blurriness of each pixel by integrating a re-blur scheme and the Walsh transform. Meanwhile, to eliminate the noise interference, we propose an adaptive sequency spectrum truncation strategy by which we can obtain an accurate blur map even in noise-polluted cases. Finally, a multi-scale fusion segmentation framework is designed to extract the blur region based on the clustering-guided region growth. Experimental results on benchmark datasets demonstrate that the proposed method achieves state-of-the-art performance and the best balance between cost and benefit. It offers an average F1 score of 0.887, MAE of 0.101, detecting time of 0.7 s, and training time of 0.5 s. Especially for noise-polluted blurry images, the proposed method achieves the F1 score of 0.887 and MAE of 0.101, which significantly surpasses other competitive approaches. Our method yields a cost–benefit advantage and noise immunity that has great application prospect in complex sensing environment.

Photon Starvation Artifact Reduction by Shift-Variant Processing

The x-ray computed tomography (CT) images with low dose are noisy and may contain photon starvation artifacts. The artifacts are location and direction dependent. Therefore, the common shift-invariant denoising filters do not work well. The state-of-the-art methods to process the low-dose CT images are image reconstruction based; they require the raw projection data. In many situations, the raw CT projections are not accessible. This paper suggests a method to denoise the low-dose CT image using the pseudo projections generated by the application of a forward projector on the low-dose CT image. The feasibility of the proposed method is demonstrated by real clinical data.

The Impact of Resampling and Denoising Deep Learning Algorithms on Radiomics in Brain Metastases MRI

Simple Summary

Due to the central role of magnetic resonance Imaging (MRI) in the management of patients with cancer, waiting lists exceed clinically relevant delays. For this reason, many research groups and MRI manufacturers develop algorithms as resampling and denoising models to allow faster acquisition time without deterioration in image quality. Whereas these algorithms are available in all new MRI, it is not clear how they will impact image features as well as the validity of statistical model of radiomics which use deep images characteristics to predict treatment outcome. The aim of this study was to develop resampling and denoising deep learning (DL) models and evaluate their impact on radiomics from post-Gd-T1w-MRI brain images with brain metastases. We show that resampling and denoising DL models reconstruct low resolution and noised MRI images acquired quickly into high quality images. While fast acquisition loses most of the radiomic-features and invalidates predictive radiomic models, DL models restore these parameters.

Abstract

Background: Magnetic resonance imaging (MRI) is predominant in the therapeutic management of cancer patients, unfortunately, patients have to wait a long time to get an appointment for examination. Therefore, new MRI devices include deep-learning (DL) solutions to save acquisition time. However, the impact of these algorithms on intensity and texture parameters has been poorly studied. The aim of this study was to evaluate the impact of resampling and denoising DL models on radiomics. Methods: Resampling and denoising DL model was developed on 14,243 T1 brain images from 1.5T-MRI. Radiomics were extracted from 40 brain metastases from 11 patients (2049 images). A total of 104 texture features of DL images were compared to original images with paired t-test, Pearson correlation and concordance-correlation-coefficient (CCC). Results: When two times shorter image acquisition shows strong disparities with the originals concerning the radiomics, with significant differences and loss of correlation of 79.81% and 48.08%, respectively. Interestingly, DL models restore textures with 46.15% of unstable parameters and 25.96% of low CCC and without difference for the first-order intensity parameters. Conclusions: Resampling and denoising DL models reconstruct low resolution and noised MRI images acquired quickly into high quality images. While fast MRI acquisition loses most of the radiomic features, DL models restore these parameters.

Low Dose CT Image Denoising Using Boosting Attention Fusion GAN with Perceptual Loss

Image denoising of Low-dose computed tomography (LDCT) images has continues to receive attention in the research community due to ongoing concerns about high-dose radiation exposure of patients for diagnosis. The use of low radiation CT image, however, could lead to inaccurate diagnosis due to the presence of noise. Deep learning techniques are being integrated into denoising methods to address this problem. In this paper, a General Adversarial Network (GAN) composed of boosting fusion of spatial and channel attention modules is proposed. These modules are embedded in the denoiser to address the limitations of other GAN-based denoising models that tend to only focus on the local processing and neglect the dependencies of creating feature maps with spatial- and channel- wise image characteristics. This study aims to preserve structural details of LDCT images by applying boosting attention modules, prevents edge over-smoothing by integrating perceptual loss via VGG16 pre-trained network, and finally, improves the computational efficiency by taking advantage of deep learning techniques and GPU parallel computation.

Five-second STEM dislocation tomography for 300 nm thick specimen assisted by deep-learning-based noise filtering

Scanning transmission electron microscopy (STEM) is suitable for visualizing the inside of a relatively thick specimen than the conventional transmission electron microscopy, whose resolution is limited by the chromatic aberration of image forming lenses, and thus, the STEM mode has been employed frequently for computed electron tomography based three-dimensional (3D) structural characterization and combined with analytical methods such as annular dark field imaging or spectroscopies. However, the image quality of STEM is severely suffered by noise or artifacts especially when rapid imaging, in the order of millisecond per frame or faster, is pursued. Here we demonstrate a deep-learning-assisted rapid STEM tomography, which visualizes 3D dislocation arrangement only within five-second acquisition of all the tilt-series images even in a 300 nm thick steel specimen. The developed method offers a new platform for various in situ or operando 3D microanalyses in which dealing with relatively thick specimens or covering media like liquid cells are required.

Imaging Dynamic Expiration: Feasibility of MRI Spirometry Using Hyperpolarized Xenon Gas

Purpose

To examine the feasibility of imaging-based spirometry using high-temporal-resolution projection MRI and hyperpolarized xenon 129 (¹²⁹Xe) gas.

Materials and Methods

In this prospective exploratory study, five healthy participants (age range, 25-45 years; three men) underwent an MRI spirometry technique using inhaled hyperpolarized ¹²⁹Xe and rapid two-dimensional projection MRI. Participants inhaled ¹²⁹Xe, then performed a forced expiratory maneuver while in an MR imager. Images of the lungs during expiration were captured in time intervals as short as 250 msec. Volume-corrected images of the lungs at expiration commencement (0 second), 1 second after expiration, and 6 seconds after expiration were extracted to generate forced expiratory volume in 1 second (FEV₁), forced vital capacity (FVC), and FEV₁/FVC ratio pulmonary maps. For comparison, participants performed conventional spirometry in the sitting position using room air, in the supine position using room air, and in the supine position using a room air and ¹²⁹Xe mixture. Paired t tests with Bonferroni corrections for multiple comparisons were used for statistical analyses.

Results

The mean MRI-derived FEV₁/FVC value was lower in comparison with conventional spirometry (0.52 ± 0.03 vs 0.70 ± 0.05, P < .01), which may reflect selective ¹²⁹Xe retention. A secondary finding of this study was that 1 L of inhaled ¹²⁹Xe negatively impacted pulmonary function as measured by conventional spirometry (in supine position), which reduced measured FEV₁ (2.70 ± 0.90 vs 3.04 ± 0.85, P < .01) and FEV₁/FVC (0.70 ± 0.05 vs 0.79 ± 0.04, P < .01).

Conclusion

A forced expiratory maneuver was successfully imaged with hyperpolarized ¹²⁹Xe and high-temporal-resolution MRI. Derivation of regional lung spirometric maps was feasible.Keywords: MR-Imaging, MR-Dynamic Contrast Enhanced, MR-Functional Imaging, Pulmonary, Thorax, Diaphragm, Lung, Pleura, Physics Supplemental material is available for this article. © RSNA, 2021.

Ring artifact reduction via multiscale nonlocal collaborative filtering of spatially correlated noise

X-ray micro-tomography systems often suffer severe ring artifacts in reconstructed images. These artifacts are caused by defects in the detector, calibration errors, and fluctuations producing streak noise in the raw sinogram data. In this work, these streaks are modeled in the sinogram domain as additive stationary correlated noise upon logarithmic transformation. Based on this model, a streak removal procedure is proposed where the Block-Matching and 3-D (BM3D) filtering algorithm is applied across multiple scales, achieving state-of-the-art performance in both real and simulated data. Specifically, the proposed fully automatic procedure allows for attenuation of streak noise and the corresponding ring artifacts without creating major distortions common to other streak removal algorithms.

Boosting the signal-to-noise of low-field MRI with deep learning image reconstruction

Recent years have seen a resurgence of interest in inexpensive low magnetic field (< 0.3 T) MRI systems mainly due to advances in magnet, coil and gradient set designs. Most of these advances have focused on improving hardware and signal acquisition strategies, and far less on the use of advanced image reconstruction methods to improve attainable image quality at low field. We describe here the use of our end-to-end deep neural network approach (AUTOMAP) to improve the image quality of highly noise-corrupted low-field MRI data. We compare the performance of this approach to two additional state-of-the-art denoising pipelines. We find that AUTOMAP improves image reconstruction of data acquired on two very different low-field MRI systems: human brain data acquired at 6.5 mT, and plant root data acquired at 47 mT, demonstrating SNR gains above Fourier reconstruction by factors of 1.5- to 4.5-fold, and 3-fold, respectively. In these applications, AUTOMAP outperformed two different contemporary image-based denoising algorithms, and suppressed noise-like spike artifacts in the reconstructed images. The impact of domain-specific training corpora on the reconstruction performance is discussed. The AUTOMAP approach to image reconstruction will enable significant image quality improvements at low-field, especially in highly noise-corrupted environments.

Denoising PET images for proton therapy using a residual U-net

The use of proton therapy has the advantage of high dose concentration as it is possible to concentrate the dose on the tumor while suppressing damage to the surrounding normal organs. However, the range uncertainty significantly affects the actual dose distribution in the vicinity of the proton range, limiting the benefit of proton therapy for reducing the dose to normal organs. By measuring the annihilation gamma rays from the produced positron emitters, it is possible to obtain a proton induced positron emission tomography (pPET) image according to the irradiation region of the proton beam. Smoothing with a Gaussian filter is generally used to denoise PET images; however, this approach lowers the spatial resolution. Furthermore, other conventional smoothing processing methods may deteriorate the steep region of the pPET images. In this study, we proposed a denoising method based on a Residual U-Net for pPET images. We conducted the Monte Carlo simulation and irradiation experiment on a human phantom to obtain pPET data. The accuracy of the range estimation and the image similarity were evaluated for pPET images using the Residual U-Net, a Gaussian filter, a median filter, the block-matching and 3D-filtering (BM3D), and a total variation (TV) filter. Usage of the Residual U-Net yielded effective results corresponding to the range estimation; however, the results of peak-signal-to-noise ratio were identical to those for the Gaussian filter, median filter, BM3D, and TV filter. The proposed method can contribute to improving the accuracy of treatment verification and shortening the PET measurement time.

A Photogrammetry-Based Workflow for the Accurate 3D Construction and Visualization of Museums Assets

Nowadays digital replicas of artefacts belonging to the Cultural Heritage (CH) are one of the most promising innovations for museums exhibitions, since they foster new forms of interaction with collections, at different scales. However, practical digitization is still a complex task dedicated to specialized operators. Due to these premises, this paper introduces a novel approach to support non-experts working in museums with robust, easy-to-use workflows based on low-cost widespread devices, aimed at the study, classification, preservation, communication and restoration of CH artefacts. The proposed methodology introduces an automated combination of acquisition, based on mobile equipment and visualization, based on Real-Time Rendering. After the description of devices used along the workflow, the paper focuses on image pre-processing and geometry processing techniques adopted to generate accurate 3D models from photographs. Assessment criteria for the developed process evaluation are illustrated. Tests of the methodology on some effective museum case studies are presented and discussed.

Deep Learning-Based Denoising for High b-Value at 2000 s/mm2 Diffusion-Weighted Imaging

Diffusion-weighted imaging (DWI) allows white matter quantification of the white matter tracts of the brain. However, at a high b-value (≥ 2000 s/mm2), DWI acquisition suffers from noise due to longer acquisition times obscuring white matter interpretation. DWI denoising techniques can be used to acquire high b-value DWI without increasing the number of signal averages. We used a residual learning-based convolutional neural network (DnCNN) to reduce noise in high b-value DWI based on the literature review. We applied the proposed denoising method on high b-value, retrospectively collected DWI data with multiple noise levels. Experimental results show an improved image quality after denoising in retrospective DWI (average PSNR before and after denoising: 27.63 ± 1.06 dB and 51.76 ± 1.95 dB, respectively). The prospective DWI included one and two signal averages for denoising. DWI with four signal averages was used as the reference. Representative images show high b-value prospective DW images denoised using the DnCNN. We demonstrated DnCNN for cases of multiple noise levels and signal averages. For the prospective study, the PSNR values for 1-NEX before and after denoising were 27.39 ± 3.75 dB and 27.68 ± 3.75 dB. For 2-NEX, the PSNR values before and after denoising were 27.51 ± 4.18 dB and 27.75 ± 4.05 dB.