Lower Image Noise Research Articles

Performance evaluation of deep learning image reconstruction algorithm for dual-energy spectral CT imaging: A phantom study.

To evaluate the performance of deep learning image reconstruction (DLIR) algorithm in dual-energy spectral CT (DEsCT) as a function of radiation dose and image energy level, in comparison with filtered-back-projection (FBP) and adaptive statistical iterative reconstruction-V (ASIR-V) algorithms. An ACR464 phantom was scanned with DEsCT at four dose levels (3.5 mGy, 5 mGy, 7.5 mGy, and 10 mGy). Virtual monochromatic images were reconstructed at five energy levels (40 keV, 50 keV, 68 keV, 74 keV, and 140 keV) using FBP, 50% and 100% ASIR-V, DLIR at low (DLIR-L), medium (DLIR-M), and high (DLIR-H) settings. The noise power spectrum (NPS), task-based transfer function (TTF) and detectability index (d') were computed and compared among reconstructions. NPS area and noise increased as keV decreased, with DLIR having slower increase than FBP and ASIR-V, and DLIR-H having the lowest values. DLIR had the best 40 keV/140 keV noise ratio at various energy levels, DLIR showed higher TTF (50%) than ASIR-V for all materials, especially for the soft tissue-like polystyrene insert, and DLIR-M and DLIR-H provided higher d' than DLIR-L, ASIR-V and FBP in all dose and energy levels. As keV increases, d' increased for acrylic insert, and d' of the 50 keV DLIR-M and DLIR-H images at 3.5 mGy (7.39 and 8.79, respectively) were higher than that (7.20) of the 50 keV ASIR-V50% images at 10 mGy. DLIR provides better noise containment for low keV images in DEsCT and higher TTF(50%) for the polystyrene insert over ASIR-V. DLIR-H has the lowest image noise and highest detectability in all dose and energy levels. DEsCT 50 keV images with DLIR-M and DLIR-H show potential for 65% dose reduction over ASIR-V50% withhigher d'.

Comparison of ultrahigh and standard resolution photon-counting CT angiography of the femoral arteries in a continuously perfused in vitro model

BackgroundWith the emergence of photon-counting CT, ultrahigh-resolution (UHR) imaging can be performed without dose penalty. This study aims to directly compare the image quality of UHR and standard resolution (SR) scan mode in femoral artery angiographies.MethodsAfter establishing continuous extracorporeal perfusion in four fresh-frozen cadaveric specimens, photon-counting CT angiographies were performed with a radiation dose of 5 mGy and tube voltage of 120 kV in both SR and UHR mode. Images were reconstructed with dedicated convolution kernels (soft: Body-vascular (Bv)48; sharp: Bv60; ultrasharp: Bv76). Six radiologists evaluated the image quality by means of a pairwise forced-choice comparison tool. Kendall’s concordance coefficient (W) was calculated to quantify interrater agreement. Image quality was further assessed by measuring intraluminal attenuation and image noise as well as by calculating signal-to-noise ratio (SNR) and contrast-to-noise ratios (CNR).ResultsUHR yielded lower noise than SR for identical reconstructions with kernels ≥ Bv60 (p < 0.001). UHR scans exhibited lower intraluminal attenuation compared to SR (Bv60: 406.4 ± 25.1 versus 418.1 ± 30.1 HU; p < 0.001). Irrespective of scan mode, SNR and CNR decreased while noise increased with sharper kernels but UHR scans were objectively superior to SR nonetheless (Bv60: SNR 25.9 ± 6.4 versus 20.9 ± 5.3; CNR 22.7 ± 5.8 versus 18.4 ± 4.8; p < 0.001). Notably, UHR scans were preferred in subjective assessment when images were reconstructed with the ultrasharp Bv76 kernel, whereas SR was rated superior for Bv60. Interrater agreement was high (W = 0.935).ConclusionsCombinations of UHR scan mode and ultrasharp convolution kernel are able to exploit the full image quality potential in photon-counting CT angiography of the femoral arteries.Relevance statementThe UHR scan mode offers improved image quality and may increase diagnostic accuracy in CT angiography of the peripheral arterial runoff when optimized reconstruction parameters are chosen.Key points• UHR photon-counting CT improves image quality in combination with ultrasharp convolution kernels.• UHR datasets display lower image noise compared with identically reconstructed standard resolution scans.• Scans in UHR mode show decreased intraluminal attenuation compared with standard resolution imaging.Graphical

Design of Low-Noise CMOS Image Sensor Using a Hybrid-Correlated Multiple Sampling Technique

We present a 320 × 240 CMOS image sensor (CIS) using the proposed hybrid-correlated multiple sampling (HMS) technique with an adaptive dual-gain analog-to-digital converter (ADC). The proposed HMS improves the noise characteristics under low illumination by adjusting the ADC gain according to the incident light on the pixels. Depending on whether it is less than or greater than 1/4 of the full output voltage range from pixels, either correlated multiple sampling or conventional-correlated double sampling (CDS) is used with different slopes of the ramping signals. The proposed CIS achieves 11-bit resolution of the ADC using an up-down counter that controls the LSB depending on the ramping signals used. The sensor was fabricated using a 0.11 μm CIS process, and the total chip area was 2.55 mm × 4.3 mm. Compared to the conventional CDS, the measurement results showed that the maximum dark random noise was reduced by 26.7% with the proposed HMS, and the maximum figure of merit was improved by 49.1%. The total power consumption was 5.1 mW at 19 frames per second with analog, pixel, and digital supply voltages of 3.3 V, 3.3 V, and 1.5 V, respectively.

A star identification algorithm for rolling shutter exposure based on Hough transform

The star identification algorithm usually identifies stars by angular distance matching. However, under high dynamic conditions, the rolling shutter effect distorts the angular distances between the measured and true star positions, leading to plethoric false matches and requiring complex and time-consuming verification for star identification. Low identification rate hinders the application of low-noise and cost-effective rolling shutter image sensors. In this work, we first study a rolling shutter distortion model of angular distances between stars, and then propose a novel three-stage star identification algorithm to identify distorted star images captured by the rolling shutter star sensor. The first stage uses a modified grid algorithm with adaptive error tolerance and an expanded pattern database to efficiently eliminate spurious matches. The second stage performs angular velocity estimation based on Hough transform to verify the matches that follow the same distortion pattern. The third stage applies a rolling shutter error correction method for further verification. Both the simulation and night sky image test demonstrate the effectiveness and efficiency of our algorithm under high dynamic conditions. The accuracy of angular velocity estimation method by Hough transform is evaluated and the root mean square error is below 0.5 (°)/s. Our algorithm achieves a 95.7% identification rate at an angular velocity of 10 (°)/s, which is much higher than traditional algorithms.

End to end multifunctional autoencoder-based classification and denoising network

Brain tumours are one of the common diseases in human beings. Currently, brain Nuclear Magnetic Resonance (MRI) is the main means of detecting brain diseases. There are many works related to the classification and noise reduction of brain MRI images, but there are few works that combine these two parts simultaneously, i.e., noise reduction processing and classification of brain images at the same time. In this work, these two parts of work will be combined together, extracting the features of brain MRI images through an encoder composed of convolutional neural networks, using the features to classify the images, and then using the features to reduce the noise of the images to produce a low-noise image. According to the model, this work also proposes a new loss function. The new loss function is composed by adjusting the weights of the loss functions related to the classification and denoising tasks. This paper tries three different ways of weight assignment, and the experimental results show that the dynamic parameter assignment approach achieves the best results for image classification, but none of the three approaches achieves acceptable results for noise reduction.

GAN-based super-resolution for confocal superficial eyelid imaging: Real-time, domain generalization, and noise robustness

Objective:Infection by Demodex mites is a significant factor in causing blepharitis. The collected superficial eyelid tissue images currently have low resolution and strong noise. High-quality reconstruction of low-quality images is needed to assist ophthalmologists in accurately and quickly diagnosing mite infestations. Methods:We propose a novel GAN-based super-resolution reconstruction model. We utilize Residual-in-Residual Dense Blocks (RRDB) as the fundamental building units and employ high-order degradation to simulate real-world degradation. Additionally, we introduce a new attention mechanism, All Attention Mechanism (AAM), which effectively captures and utilizes crucial image features, expanding the scope of utilized pixels. This significantly enhances the quality of reconstructed images. Results:Compared to existing super-resolution reconstruction models, our model demonstrates the strongest generalization ability and robustness. On the CIDMG and FMD datasets, we achieve the lowest NIQE and the highest SSIM values. Our model outperforms the SOTA SwinIR model, reducing reconstruction time by 76% and improving the image quality by 6.6% for low-noise images and 15.3% for high-noise images. Conclusion:By combining GAN with AAM, we have significantly improved the model’s generalization ability, robustness, and real-time performance. Significance:This is the first time that domain generalization has been used for reconstructing superficial eyelid tissue images. The stable and fast reconstruction performance reduces the impact of noise and low resolution on image quality, significantly reducing examination time and improving diagnostic accuracy. This greatly enhances the clinical utility of confocal microscopy for eyelid demodex examination.

A novel multi-agent genetic algorithm for limited-view computed tomography

Computed tomography (CT) is a non-invasive testing technique to generate the internal image of any object. It has various applications belonging to medical imaging and other engineering applications. The present article proposes a novel multiagent genetic algorithm based on hill-climbing for limited-view computed tomography (MAGAH-CT). The MAGAH combines a multi-agent system based on a genetic algorithm (MAGA) and a hill-climbing approach. The MAGAH uses multiple operators to refine the agent population during evolution. These operators are neighborhood-based competition, hybrid crossover, adaptive mutation, and self-learning. MAGAH-CT explores better solutions in neighborhood-based competition and cooperation operators. The self-learning operator learns through the agent’s population generated by adaptive mutation and hybrid crossover operator, ensuring the results’ continuity. The proposed algorithm has been tested on a Shepp-Logan head phantom with limited view projection data. The algorithm can accurately reconstruct the low-noise CT image using limited data constraints. Experimental results reveal that the presented MAGAH-CT algorithm is efficient and computationally reasonable. The presented algorithm is suitable for low-noise CT reconstruction and out-performs other limited-view CT reconstruction algorithms.

U-shaped fusion convolutional transformer based workflow for fast optical coherence tomography angiography generation in lips.

Oral disorders, including oral cancer, pose substantial diagnostic challenges due to late-stage diagnosis, invasive biopsy procedures, and the limitations of existing non-invasive imaging techniques. Optical coherence tomography angiography (OCTA) shows potential in delivering non-invasive, real-time, high-resolution vasculature images. However, the quality of OCTA images are often compromised due to motion artifacts and noise, necessitating more robust and reliable image reconstruction approaches. To address these issues, we propose a novel model, a U-shaped fusion convolutional transformer (UFCT), for the reconstruction of high-quality, low-noise OCTA images from two-repeated OCT scans. UFCT integrates the strengths of convolutional neural networks (CNNs) and transformers, proficiently capturing both local and global image features. According to the qualitative and quantitative analysis in normal and pathological conditions, the performance of the proposed pipeline outperforms that of the traditional OCTA generation methods when only two repeated B-scans are performed. We further provide a comparative study with various CNN and transformer models and conduct ablation studies to validate the effectiveness of our proposed strategies. Based on the results, the UFCT model holds the potential to significantly enhance clinical workflow in oral medicine by facilitating early detection, reducing the need for invasive procedures, and improving overall patient outcomes.

Objective Image Quality Optimization in Augmented Reality Using Spatial Frequency Domain Models.

Augmented reality (AR) blends the digital and physical worlds by overlapping a virtual image onto the see-through physical environment. However, contrast reduction and noise superposition in an AR head-mounted display (HMD) can substantially limit image quality and human perceptual performance in both the digital and physical spaces. To assess image quality in AR, we performed human and model observer studies for various imaging tasks with targets placed in the digital and physical worlds. A target detection model was developed for the complete AR system including the optical see-through. Target detection performance using different observer models developed in the spatial frequency domain was compared with the human observer results. The non-prewhitening model with eye filter and internal noise results closely track human perception performance as measured by the area under the receiver operating characteristic curve (AUC), especially for tasks with high image noise. The AR HMD non-uniformity limits the low-contrast target (less than 0.02) observer performance for low image noise. In augmented reality conditions, the detectability of a target in the physical world is reduced due to the contrast reduction by the overlaid AR display image (AUC less than 0.87 for all the contrast levels evaluated). We propose an image quality optimization scheme to optimize the AR display configurations to match observer detection performance for targets in both the digital and physical worlds. The image quality optimization procedure is validated using both simulation and bench measurements of a chest radiography image with digital and physical targets for various imaging configurations.

Validation of deep learning-based CT image reconstruction for treatment planning

Deep learning-based CT image reconstruction (DLR) is a state-of-the-art method for obtaining CT images. This study aimed to evaluate the usefulness of DLR in radiotherapy. Data were acquired using a large-bore CT system and an electron density phantom for radiotherapy. We compared the CT values, image noise, and CT value-to-electron density conversion table of DLR and hybrid iterative reconstruction (H-IR) for various doses. Further, we evaluated three DLR reconstruction strength patterns (Mild, Standard, and Strong). The variations of CT values of DLR and H-IR were large at low doses, and the difference in average CT values was insignificant with less than 10 HU at doses of 100 mAs and above. DLR showed less change in CT values and smaller image noise relative to H-IR. The noise-reduction effect was particularly large in the low-dose region. The difference in image noise between DLR Mild and Standard/Strong was large, suggesting the usefulness of reconstruction intensities higher than Mild. DLR showed stable CT values and low image noise for various materials, even at low doses; particularly for Standard or Strong, the reduction in image noise was significant. These findings indicate the usefulness of DLR in treatment planning using large-bore CT systems.

Deep learning-based iodine contrast-augmenting algorithm for low-contrast-dose liver CT to assess hypovascular hepatic metastasis.

To investigate the image quality and diagnostic performance of low-contrast-dose liver CT using a deep learning-based iodine contrast-augmenting algorithm (DLICA) for hypovascular hepatic metastases. This retrospective study included 128 patients who underwent contrast-enhanced dual-energy CT for hepatic metastasis surveillance between July 2019 and June 2022 using a 30% reduced iodine contrast dose in the portal phase. Three image types were reconstructed: 50-keV virtual monoenergetic images (50-keV VMI); linearly blended images simulating 120-kVp images (120-kVp); and post-processed 120-kVp images using DLICA (DLICA 120-kVp). Three reviewers evaluated lesion conspicuity, image contrast, and subjective image noise. We also measured image noise, contrast-to-noise ratios (CNRs), and signal-to-noise ratios (SNRs). The diagnostic performance for hepatic metastases was evaluated using a jackknife alternative free-response receiver operating characteristic method with the consensus of two independent radiologists as the reference standard. DLICA 120-kVp demonstrated significantly higher CNR of lesions to liver (5.7±3.1 vs. 3.8±2.1 vs. 3.8±2.1) and higher SNR compared with 50-keV VMI and 120-kVp (p<0.001 for all). DLICA 120-kVp had significantly lower image noise than 50-kVp VMI for all regions (p<0.001 for all). DLICA 120-kVp also exhibited superior lesion conspicuity (4.0 [3.3-4.3] vs. 3.7 [3.0-4.0] vs. 3.7 [3.0-4.0]), higher image contrast, and lower subjective image noise compared with 50-keV VMI and 120-kVp (p<0.001 for all). Although there was no significant difference in the figure of merit for lesion diagnosis among the three methods (p=0.11), DLICA 120-kVp had a significantly higher figure of merit for lesions with a diameter <20 mm than 50-keV VMI (0.677 vs. 0.648, p=0.007). On a per-lesion basis, DLICA 120-kVp also demonstrated higher sensitivity than the 50-keV VMI (81.2% vs. 72.9%, p<0.001). The specificities per lesion were not significantly different among the three algorithms (p=0.15). DLICA at 120-kVp provided superior lesion conspicuity and image quality and similar diagnostic performance for hypovascular hepatic metastases compared with 50-keV VMI.

Deep learning reconstruction vs standard reconstruction for abdominal CT: the influence of BMI.

This study aimed to evaluate the image quality and lesion conspicuity of the deep learning image reconstruction (DLIR) algorithm compared with standard image reconstruction algorithms on abdominal enhanced computed tomography (CT) scanning with a wide range of body mass indexes (BMIs). A total of 112 participants who underwent contrast-enhanced abdominal CT scans were divided into three groups according to BMIs: the 80-kVp group (BMI ≤ 23.9kg/m2), 100-kVp group (BMI 24-28.9kg/m2), and 120-kVp group (BMI ≥ 29kg/m2). All images were reconstructed using filtered back projection (FBP), adaptive statistical iterative reconstruction-V of 50% level (IR), and DLIR at low, medium, and high levels (DL, DM, and DH, respectively). Subjective noise, artifact, overall image quality, and low- and high-contrast hepatic lesion conspicuity were all graded on a 5-point scale. The CT attenuation value (in HU), image noise, and contrast-to-noise ratio (CNR) were quantified and compared. DM and DH improved the qualitative and quantitative parameters compared with FBP and IR for all three BMI groups. DH had the lowest image noise and highest CNR value, while DM had the highest subjective overall image quality and low- and high-contrast lesion conspicuity scores for the three BMI groups. Based on the FBP, the improvement in image quality and lesion conspicuity of DM and DH images was greater in the 80-kVp group than in the 100-kVp and 120-kVp groups. For all BMIs, DLIR improves both image quality and hepatic lesion conspicuity, of which DM would be the best choice to balance both. The study suggests that utilizing DLIR, particularly at the medium level, can significantly enhance image quality and lesion visibility on abdominal CT scans across a wide range of BMIs. • DLIR improved the image quality and lesion conspicuity across a wide range of BMIs. • DLIR at medium level had the highest subjective parameters and lesion conspicuity scores among all reconstruction levels. • On the basis of the FBP, the 80-kVp group had improved image quality and lesion conspicuity more than the 100-kVp and 120-kVp groups.

Long Versus Short Axial Field of View Immuno-PET/CT: Semiquantitative Evaluation for 89Zr-Trastuzumab.

The purpose of this study was to quantify any differences between the SUVs of 89Zr immuno-PET scans obtained using a PET/CT system with a long axial field of view (LAFOV; Biograph Vision Quadra) compared to a PET/CT system with a short axial field of view (SAFOV; Biograph Vision) and to evaluate how LAFOV PET scan duration affects image noise and SUV metrics. Methods: Five metastatic breast cancer patients were scanned consecutively on SAFOV and LAFOV PET/CT scanners. Four additional patients were scanned using only LAFOV PET/CT. Scans on both systems lasted approximately 30 min and were acquired 4 d after injection of 37 MBq of 89Zr-trastuzumab. LAFOV list-mode data were reprocessed to obtain images acquired using shorter scan durations (15, 10, 7.5, 5, and 3 min). Volumes of interest were placed in healthy tissues, and tumors were segmented semiautomatically to compare coefficients of variation and to perform Bland-Altman analysis on SUV metrics (SUVmax, SUVpeak, and SUVmean). Results: Using 30-min images, 2 commonly used lesion SUV metrics were higher for SAFOV than for LAFOV PET (SUVmax, 16.2% ± 13.4%, and SUVpeak, 10.1% ± 7.2%), whereas the SUVmean of healthy tissues showed minimal differences (0.7% ± 5.8%). Coefficients of variation in the liver derived from 30-min SAFOV PET were between those of 3- and 5-min LAFOV PET. The smallest SUVmax and SUVpeak differences between SAFOV and LAFOV were found for 3-min LAFOV PET. Conclusion: LAFOV 89Zr immuno-PET showed a lower SUVmax and SUVpeak than SAFOV because of lower image noise. LAFOV PET scan duration may be reduced at the expense of increasing image noise and bias in SUV metrics. Nevertheless, SUVpeak showed only minimal bias when reducing scan duration from 30 to 10 min.

DICOM file format has better radiographic image quality than other file formats: an objective study.

This study aimed to assess the influence of the file format on the image quality parameters (image noise, brightness, and uniformity) of periapical radiographs acquired with different digital systems. Radiographic images of an acrylic phantom were acquired with two digital systems - Digora Toto and Express, and exported into five different file formats - TIFF, BMP, DICOM, PNG, and JPEG. Image noise, image brightness (mean of gray values), and image uniformity (standard deviation of gray values) were evaluated in all images. A two-way analysis of variance with Tukey's test as a post-hoc test was used to compare the results, considering the file formats and radiographic systems as the studied factors. A significance level of 5% was adopted for all analyses. The DICOM image file format presented lower image noise, higher brightness (higher mean gray values), and greater image uniformity (p<0.001) than the other file formats, which did not differ from each other for both digital radiography systems (p>0.05). The Express system revealed lower image noise and greater image uniformity than the Digora Toto system regardless of the image file format (p<0.001). Moreover, the Express showed higher brightness than the Digora Toto for all image file formats (p<0.001), except for the DICOM image file format, which did not significantly differ between the digital radiography systems tested (p>0.05). The DICOM image file format showed lower image noise, higher brightness, and greater image uniformity than the other file formats (TIFF, BMP, PNG, and JPEG) in both digital radiography systems tested.

Image Quality and Lesion Detectability of Pancreatic Phase Thin-Slice Computed Tomography Images With a Deep Learning-Based Reconstruction Algorithm.

To evaluate the image quality and lesion detectability of pancreatic phase thin-slice computed tomography (CT) images reconstructed with a deep learning-based reconstruction (DLR) algorithm compared with filtered-back projection (FBP) and hybrid iterative reconstruction (IR) algorithms. Fifty-three patients who underwent dynamic contrast-enhanced CT including pancreatic phase were enrolled in this retrospective study. Pancreatic phase thin-slice (0.625 mm) images were reconstructed with each FBP, hybrid IR, and DLR. Objective image quality and signal-to-noise ratio of the pancreatic parenchyma, and contrast-to-noise ratio of pancreatic lesions were compared between the 3 reconstruction algorithms. Two radiologists independently assessed the image quality of all images. The diagnostic performance for the detection of pancreatic lesions was compared among the reconstruction algorithms using jackknife alternative free-response receiver operating characteristic analysis. Deep learning-based reconstruction resulted in significantly lower image noise and higher signal-to-noise ratio and contrast-to-noise ratio than hybrid IR and FBP ( P < 0.001). Deep learning-based reconstruction also yielded significantly higher visual scores than hybrid IR and FBP ( P < 0.01). The diagnostic performance of DLR for detecting pancreatic lesions was highest for both readers, although a significant difference was found only between DLR and FBP in one reader ( P = 0.02). Deep learning-based reconstruction showed improved objective and subjective image quality of pancreatic phase thin-slice CT relative to other reconstruction algorithms and has potential for improving lesion detectability.

Electric‐Dipole Gated Two Terminal Phototransistor for Charge‐Coupled Device

AbstractThe demand for charge‐coupled device (CCD) imagers has surged exponentially during the last decade owing to their exceptionally high quality and low noise imaging. However, they are still confronting the performance constraints of high operation power, low speed, and limited charge integration. Here, the electric‐dipole gated phototransistor operation without external gate bias is reported by using high‐k HfO2 dielectric material. The electrostatic coupling of photogenerated charges from the Si with the graphene channel through a 10 nm HfO2 layer is demonstrated. The device exhibits remarkable performance in the broadband spectrum (266–1342 nm) at low drain bias voltage. The high values of responsivity, external quantum efficiency, and detectivity of 3.7 × 103 A W−1, 0.72 × 104, and 6.20 × 1013 cmHz½ W−1, respectively, for 800 nm wavelength and 3.3 × 103 A W−1, 1.31 × 104, and 5.61 × 1013 cmHz½ W−1, respectively, for 400 nm wavelength without gate are achieved. This discovery may potentially eliminate the requirement for gate terminals from commercial CCD devices. The power efficient features of this gateless image sensor can be fabricated at the industrial scale for the future machine vision market.

Gaussian-weighted crosscorrelation imaging condition for microseismic source localization

Time-reversal imaging is a powerful method for the localization of microseismic events. Conventional time-reversal imaging methods such as autocorrelation imaging or grouped crosscorrelation imaging may suffer from imaging artifacts, e.g., caused by less refined velocity models, noise-contaminated data, and data acquired from sparse receiver networks. These artifacts typically reduce imaging quality and may cause subsequent misinterpretations leading to false positives. To address these issues, we develop a new imaging condition that comprises three steps for each time step. First, we divide the back-propagated wavefield into parts according to their maximum absolute amplitudes. Second, the amplitudes of the back-propagated wavefield are weighted by a Gaussian function with the spatial extent of the prevailing wavelength of the event, centered at the absolute maximum of that part of the wavefield. Finally, we zero-lag crosscorrelate these weighted wavefields at each space point to obtain an image for this time step. The final image is gained by summing the images for each time step. This process collects all energy concentrations along the back projection process, and the energy on the wavefront overlaps and collapses at the hypocenter leading to high-resolution images displaying little to no imaging artifacts. Numerical examples using the Marmousi-II and the 3D SEG overthrust models and a 3D field data example indicate the performance of our method. High-resolution low-noise source images allow unique identification of sources even for source clusters, noisy data, and sparse acquisitions. The source localization errors are smaller than the dominant wavelength of the signal, where a smooth model with a mean velocity error of approximately 5% was considered in the synthetic examples, and a homogeneous model was used in the field data example.

Improving image quality with super-resolution deep-learning-based reconstruction in coronary CT angiography.

To evaluate the effect of super-resolution deep-learning-based reconstruction (SR-DLR) on the image quality of coronary CT angiography (CCTA). Forty-one patients who underwent CCTA using a 320-row scanner were retrospectively included. Images were reconstructed with hybrid (HIR), model-based iterative reconstruction (MBIR), normal-resolution deep-learning-based reconstruction (NR-DLR), and SR-DLR algorithms. For each image series, image noise, and contrast-to-noise ratio (CNR) at the left main trunk, right coronary artery, left anterior descending artery, and left circumflex artery were quantified. Blooming artifacts from calcified plaques were measured. Image sharpness, noise magnitude, noise texture, edge smoothness, overall quality, and delineation of the coronary wall, calcified and noncalcified plaques, cardiac muscle, and valves were subjectively ranked on a 4-point scale (1, worst; 4, best). The quantitative parameters and subjective scores were compared among the four reconstructions. Task-based image quality was assessed with a physical evaluation phantom. The detectability index for the objects simulating the coronary lumen, calcified plaques, and noncalcified plaques was calculated from the noise power spectrum (NPS) and task-based transfer function (TTF). SR-DLR yielded significantly lower image noise and blooming artifacts with higher CNR than HIR, MBIR, and NR-DLR (all p < 0.001). The best subjective scores for all the evaluation criteria were attained with SR-DLR, with significant differences from all other reconstructions (p < 0.001). In the phantom study, SR-DLR provided the highest NPS average frequency, TTF50%, and detectability for all task objects. SR-DLR considerably improved the subjective and objective image qualities and object detectability of CCTA relative to HIR, MBIR, and NR-DLR algorithms. The novel SR-DLR algorithm has the potential to facilitate accurate assessment of coronary artery disease on CCTA by providing excellent image quality in terms of spatial resolution, noise characteristics, and object detectability. • SR-DLR designed for CCTA improved image sharpness, noise property, and delineation of cardiac structures with reduced blooming artifacts from calcified plaques relative to HIR, MBIR, and NR-DLR. • In the task-based image-quality assessments, SR-DLR yielded better spatial resolution, noise property, and detectability for objects simulating the coronary lumen, coronary calcifications, and noncalcified plaques than other reconstruction techniques. • The image reconstruction times of SR-DLR were shorter than those of MBIR, potentially serving as a novel standard-of-care reconstruction technique for CCTA performed on a 320-row CT scanner.

DIRECT, a low-cost system for high-speed, low-noise imaging of fluorescent bio-samples.

A targeted imaging system has been developed for applications requiring recording from stationary samples at high spatiotemporal resolutions. It works by illuminating regions of interest in rapid sequence, and recording the signal from the whole field of view onto a single photodetector. It can be implemented at low cost on an existing microscope without compromising existing functionality. The system is characterized in terms of speed, spatial resolution, and tissue penetration depth, before being used to record individual action potentials from ASAP-3 expressing neurons in an ex vivo mouse brain slice preparation.

Improving image quality and resolution of coronary arteries in coronary computed tomography angiography by using high-definition scans and deep learning image reconstruction.

Coronary computed tomography angiography (CTA) has been increasingly used to identify the degree of coronary artery stenosis and plaque lesions in vessels. This study evaluated the feasibility of using high-definition (HD) scanning with high-level deep learning image reconstruction (DLIR-H) to improve the image quality and spatial resolution when imaging calcified plaques and stents in coronary CTA as compared to the standard definition (SD) reconstruction mode with adaptive statistical iterative reconstruction-V (ASIR-V). A total of 34 patients (age 63.3±10.9 years; 55.88% female) with calcified plaques and/or stents who underwent coronary CTA in HD-mode were included in this study. Images were reconstructed with SD-ASIR-V, HD-ASIR-V, and HD-DLIR-H. Subjective image quality with image noise and clarity of vessels, calcifications, and stented lumens was evaluated by 2 radiologists using a 5-point scale. The kappa (κ) test was used to analyze the interobserver agreement. Objective image quality with image noise, signal-to-noise-ratio (SNR), and contrast-to-noise-ratio (CNR) was measured and compared. Image spatial resolution and beam-hardening artifacts (BHAs) were also evaluated using the calcification diameter and CT numbers in 3 points along the stented lumen (inside, at the proximal and distal ends just outside stent). There were 45 calcified plaques and 4 coronary stents. HD-DLIR-H images had the highest overall image quality score (4.50±0.63) with the lowest image noise (22.59±3.59 HU) and the highest SNR (18.30±4.88) and CNR (26.56±6.33), followed by SD-ASIR-V50% image quality score (4.06±2.49), image noise (35.02±8.09 HU), SNR (12.77±1.59), CNR(15.67±1.92) and HD-ASIR-V50% image quality score (3.90±0.64), image noise (57.7±12.03 HU), SNR (8.16±1.86), CNR (10.01±2.39). HD-DLIR-H images also had the smallest calcification diameter measurement (2.36±1.58 mm), followed by HD-ASIR-V50% (3.46±2.07 mm) and SD-ASIR-V50% (4.06±2.49 mm). HD-DLIR-H images had the closest CT value measurements for the 3 points along the stented lumen, indicating much less BHA. Interobserver agreement on the image quality assessment was good to excellent (HD-DLIR-H: κ value =0.783; HD-ASIR-V50%: κ value =0.789; SD-ASIR-V50%: κ value =0.671). Coronary CTA with HD scan mode and DLIR-H significantly improves the spatial resolution for displaying calcifications and in-stent lumens while simultaneously reducing image noise.