Image Sharpness Research Articles

Training certified detectives to track down the intrinsic shortcuts in COVID-19 chest x-ray data sets

Deep learning faces a significant challenge wherein the trained models often underperform when used with external test data sets. This issue has been attributed to spurious correlations between irrelevant features in the input data and corresponding labels. This study uses the classification of COVID-19 from chest x-ray radiographs as an example to demonstrate that the image contrast and sharpness, which are characteristics of a chest radiograph dependent on data acquisition systems and imaging parameters, can be intrinsic shortcuts that impair the model’s generalizability. The study proposes training certified shortcut detective models that meet a set of qualification criteria which can then identify these intrinsic shortcuts in a curated data set.

Learning to Super-Resolve Blurry Images With Events

Super-Resolution from a single motion Blurred image (SRB) is a severely ill-posed problem due to the joint degradation of motion blurs and low spatial resolution. In this paper, we employ events to alleviate the burden of SRB and propose an Event-enhanced SRB (E-SRB) algorithm, which can generate a sequence of sharp and clear images with High Resolution (HR) from a single blurry image with Low Resolution (LR). To achieve this end, we formulate an event-enhanced degeneration model to consider the low spatial resolution, motion blurs, and event noises simultaneously. We then build an event-enhanced Sparse Learning Network (eSL-Net++) upon a dual sparse learning scheme where both events and intensity frames are modeled with sparse representations. Furthermore, we propose an event shuffle-and-merge scheme to extend the single-frame SRB to the sequence-frame SRB without any additional training process. Experimental results on synthetic and real-world datasets show that the proposed eSL-Net++ outperforms state-of-the-art methods by a large margin. Datasets, codes, and more results are available at https://github.com/ShinyWang33/eSL-Net-Plusplus.

Tumor Detection in Brain Using MRI Images

This work provides a robust segmentation approach that combines Template-based K-means with the modified Fuzzy C-means (TKFCM) clustering algorithm. Operator and equipment error is reduced. The template is chosen in this method based on the convolution of grey level intensity in a tiny area of a brain image and a brain tumor image. The K-means method emphasizes initial segmentation through template selection. Distances from the cluster centroid to the cluster data points are used to update membership until it is optimal. This Euclidian distance is determined by the coarse image's intensity, entropy, contrast, dissimilarity, and homogeneity which were solely based on resemblance in conventional FCM. Then, using updated membership and automatic cluster selection, a sharp segmented image with red indicated tumor is created using the improved FCM technique. TKFCM detects minor differences in grey level intensity between normal and diseased tissue. The TKFCM method's performance is analyzed using a neural network to produce better regression and less error. The performance parameters yield meaningful results that are helpful at finding cancers in several different intensity-based brain scans brain. Keywords - Magnetic Resonance Imaging (MRI), Revised Fuzzy C-Means Algorithms and Template based K-Mean Clustering (TKFCM), gray level intensity, coarse image, features selection, Artificial Neural Network (ANN).

Reverse time migration for imaging periodic obstacles with electromagnetic plane wave

Abstract We propose novel reverse time migration (RTM) methods for the imaging of periodic obstacles using only measurements from the lower or upper side of the obstacle arrays at a fixed frequency. We analyze the resolution of the lower side and upper side RTM methods in terms of propagating modes of the Rayleigh expansion, Helmholtz–Kirchhoff equation and the distance of the measurement surface to the obstacle arrays, where the periodic structure leads to novel analysis. We give some numerical experiments to justify the competitive efficiency of our imaging functionals and the robustness against noises. Further, numerical experiments show sharp images especially for the vertical part of the periodic obstacle in the lower-RTM case, which is not shared by results for imaging bounded compactly supported obstacles.

Intra-individual comparison of image quality of the coronary arteries between photon-counting detector and energy-integrating detector CT systems

PurposeTo intra-individually compare the objective and subjective image quality of coronary computed tomography angiography (CCTA) between photon-counting detector CT (PCD-CT) and energy-integrating detector CT (EID-CT). MethodConsecutive patients undergoing clinically indicated CCTA on an EID-CT system were prospectively enrolled for a research CCTA performed on a PCD-CT system within 30 days. Polychromatic images were reconstructed for both EID- and PCD-CT, while virtual monoenergetic images (VMI) were generated at 40, 45, 50, 55, 60 and 70 keV for PCD-CT. Two blinded readers calculated contrast-to-noise ratio (CNR) for each major coronary artery and rated image noise, vessel attenuation, vessel sharpness, and overall quality on a 1–5 Likert scale. Patients were then stratified by body mass index (BMI) [high (>30 kg/m2) vs low (<30 kg/m2)] for subgroup analysis. ResultsA total of 20 patients (67.5 ± 9.0 years, 75% male) were included in the study. Compared with EID-CT, coronary artery CNR values from PCD-CT monoenergetic and polychromatic reconstructions were all significantly higher than CNR values from EID-CT, with incrementally greater differences in obese subjects (all p < 0.008). Subjective image noise and sharpness were also significantly higher for all VMI reconstructions compared to EID-CT (all p < 0.008). All subjective scores were significantly higher for 55, 60, and 70 keV PCD-CT than EID-CT values (all p < 0.05). ConclusionsThe improved objective and subjective image quality of PCD-CT compared to EID-CT may provide better visualization of the coronary arteries for a wide array of patients, especially those with a high BMI.

Deep Learning-Based Motion Correction in Projection Domain for Coronary Computed Tomography Angiography: A Clinical Evaluation.

This study aimed to evaluate the clinical performance of a deep learning-based motion correction algorithm (MCA) in projection domain for coronary computed tomography angiography (CCTA). A total of 192 patients who underwent CCTA examinations were included and divided into 2 groups based on the average heart rate (HR): group 1, 82 patients with HR of <75 beats per minute; group 2, 110 patients with HR of ≥75 beats per minute. The CCTA images were reconstructed with and without MCA. The subjective image quality was graded in terms of vessel visualization, sharpness, diagnostic confidence, and overall image quality using a 5-point scale, where cases with all scores of ≥3 were deemed interpretable. Objective image quality was measured through signal-to-noise ratio and contrast-to-noise ratio in regions relative to the vessels. The image quality scores for 2 reconstructions and effective dose between 2 groups were compared. The mean effective dose was similar between 2 groups. Neither group showed significant difference on objective image quality for 2 reconstructions. Images reconstructed with and without MCA were both found interpretable for group 1, whereas the subjective image quality was significantly improved by the MCA for all 4 metrics in group 2, with the interpretability increased from 80.91% to 99.09%. Compared with group 1, group 2 showed similar interpretability and diagnostic confidence, despite inferior overall image quality. In CCTA examinations, the deep learning-based MCA is capable of improving the image quality and diagnostic confidence for patients with increased HR to a similar level as for those with low HR.

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.

Multi-patch multi-scale model for motion deblurring with high-frequency information

Nowadays deep learning-based models have made encouraging accomplishments in the field of image motion deblurring, but this type of algorithms does not effectively use the known prior knowledge and the physical model of motion blur. So as to settle the above disputes, this paper suggests a multi-patch multi-scale input model that fuses high-frequency information, structural self-similarity (SSS), multi-scale channel spatial attention (MSCSA), and superpixel gradient loss for motion deblurring. The multi-patch multi-scale input model can fully extract the salient information, scale information, and high-frequency information of the image to restore the sharp image. The SSS module can take full advantage of the prior information of the structure self-similarity of the image. And the MSCSA module dynamically and selectively enhances relevant features and suppresses irrelevant features by fusing spatial, scale, and channel characteristics. Moreover, this paper advises a superpixel gradient loss function, which can remove motion blur and restore sharp images more effectively. And to increase the diversity of scenes, for the first time, we propose a motion blur dataset (Underwater-Blur) specifically for underwater scenes. Massive comparative experiments conducted on three public datasets of GoPro, RealBlur, RWBI, and the Underwater-Blur dataset prove that our model outperforms the state-of-the-art in terms of comprehensive performance.

Optimized flow compensation for contrast-enhanced T1-weighted Wave-CAIPI 3D MPRAGE imaging of the brain

Flow-related artifacts have been observed in highly accelerated T1-weighted contrast-enhanced wave-controlled aliasing in parallel imaging (CAIPI) magnetization-prepared rapid gradient-echo (MPRAGE) imaging and can lead to diagnostic uncertainty. We developed an optimized flow-mitigated Wave-CAIPI MPRAGE acquisition protocol to reduce these artifacts through testing in a custom-built flow phantom. In the phantom experiment, maximal flow artifact reduction was achieved with the combination of flow compensation gradients and radial reordered k-space acquisition and was included in the optimized sequence. Clinical evaluation of the optimized MPRAGE sequence was performed in 64 adult patients, who all underwent contrast-enhanced Wave-CAIPI MPRAGE imaging without flow-compensation and with optimized flow-compensation parameters. All images were evaluated for the presence of flow-related artifacts, signal-to-noise ratio (SNR), gray-white matter contrast, enhancing lesion contrast, and image sharpness on a 3-point Likert scale. In the 64 cases, the optimized flow mitigation protocol reduced flow-related artifacts in 89% and 94% of the cases for raters 1 and 2, respectively. SNR, gray-white matter contrast, enhancing lesion contrast, and image sharpness were rated as equivalent for standard and flow-mitigated Wave-CAIPI MPRAGE in all subjects. The optimized flow mitigation protocol successfully reduced the presence of flow-related artifacts in the majority of cases.Relevance statementAs accelerated MRI using novel encoding schemes become increasingly adopted in clinical practice, our work highlights the need to recognize and develop strategies to minimize the presence of unexpected artifacts and reduction in image quality as potential compromises to achieving short scan times.Key points• Flow-mitigation technique led to an 89–94% decrease in flow-related artifacts.• Image quality, signal-to-noise ratio, enhancing lesion conspicuity, and image sharpness were preserved with the flow mitigation technique.• Flow mitigation reduced diagnostic uncertainty in cases where flow-related artifacts mimicked enhancing lesions.Graphical

Image reconstruction method for electrical impedance tomography based on RBF and attention mechanism

Electrical impedance tomography (EIT) is an imaging technology for estimating the conductivity change for the measured field of the human body. However, because EIT image reconstruction is severely ill-posed and nonlinear inverse problem, the reconstructed image based on traditional algorithm is lower accurate with obscure boundaries and lots of artifacts. In order to improve the accuracy of reconstructed image, an EIT image reconstruction method based on RBF and attention mechanism is proposed in this paper. The proposed RCU-Net network architecture consists of two sub-networks: RBF and CBAM-UNet. The RBF neural network accomplishes the nonlinear mapping between the boundary voltages and conductivity distribution values. The CBAM-UNet combining attention mechanism CBAM with U-Net makes further processing for the resulting images to obtain more sharp and explicit images. RCU-Net was trained and tested on the simulated dataset created by EIDORS platform. The experiment reveals that the proposed method outperforms the traditional methods in distinguishing multi-target, generalization ability and anti-interference, and it has improved the reconstructed image quality which has an average reduction of 55.09% and 26.23% on RE (Relative Error) and an increase of 14.19% and 3.08% on CC (Correlation Coefficient) respectively compared with RBF-EIT and RBF-UNet methods. In addition, the experimental data from KIT4 system can be applied to the trained network. This successful transition also demonstrates the proposed method can reconstruct more explicit images with high fidelity.

An anisotropically diffusible fuzzy clustering based image segmentation methodology for intracranial cyst detection

An Intracranial cyst is an abnormal growth of mass in the brain that affects functioning of the nervous system and so an early detection of the lesion enables to avoid adverse effects. The processing unit in the Magnetic Resonance Imaging (MRI) system performs reading the images followed by primary image enhancement to suppress distortions thereby enhancing the feature quality in terms of its intensity, augmenting the resolution by image segmentation, post-processing by thresholding based on grayscale values and performing several morphological operations. With the existing methodologies, extracting the Region Of Interest (ROI) with the overlapping intensity values lead to inaccurate results. A novel method in which the input image that is anisotropically diffused and blurred is converted into a sharp image. Further, fuzzy partitioning of pixels deployed on Global Thresholding –Clustering Methodology (GT-CM) based segmentation takes 4 clusters into account hence forth seperating the exterior portion of the skull, the border region of the skull, the ventricles which may include the lesion and the noise. Statistical results based on several metrics such as sensitivity, specificity, F measure, Jaccord Index, Dice Coefficient and precision show that the proposed method is far more effective. An accuracy of 99.26% is obtained in exactly locating and extracting the lesion along with its attributes.

Free-breathing high isotropic resolution quantitative susceptibility mapping (QSM) of liver using 3D multi-echo UTE cones acquisition and respiratory motion-resolved image reconstruction.

To enable free-breathing and high isotropic resolution liver quantitative susceptibility mapping (QSM) using 3D multi-echo UTE cones acquisition and respiratory motion-resolved image reconstruction. Using 3D multi-echo UTE cones MRI, a respiratory motion was estimated from the k-space center of the imaging data. After sorting the k-space data with estimated motion, respiratory motion state-resolved reconstruction was performed for multi-echo data followed by nonlinear least-squares fitting for proton density fat fraction (PDFF), , and fat-corrected B0 field maps. PDFF and B0 field maps were subsequently used for QSM reconstruction. The proposed method was compared with motion-averaged (gridding) reconstruction and conventional 3D multi-echo Cartesian MRI in moving gadolinium phantom and in vivo studies. Region of interest (ROI)-based linear regression analysis was performed on these methods to investigate correlations between gadolinium concentration and QSM in the phantom study and between and QSM in in vivo study. Cones with motion-resolved reconstruction showed sharper image quality compared to motion-averaged reconstruction with a substantial reduction of motion artifacts in both moving phantom and in vivo studies. For ROI-based linear regression analysis of the phantom study, susceptibility values from cones with motion-resolved reconstruction ( = 0.31 0.05, = 0.999) and Cartesian without motion ( = 0.32 0.04, = 1.000) showed linear relationships with gadolinium concentrations and showed good agreement with each other. For in vivo, motion-resolved reconstruction showed higher goodness of fit ( = 0.00261 0.524, = 0.977) compared to motion-averaged reconstruction ( = 0.0021 0.572, = 0.723) in ROI-based linear regression analysis between and QSM. Feasibility of free-breathing liver QSM was demonstrated with motion-resolved 3D multi-echo UTE cones MRI, achieving high isotropic resolution currently unachievable in conventional Cartesian MRI.

Real Image Deblurring Based on Implicit Degradation Representations and Reblur Estimation

Most existing image deblurring methods are based on the estimation of blur kernels and end-to-end learning of the mapping relationship between blurred and sharp images. However, since different real-world blurred images typically have completely different blurring patterns, the performance of these methods in real image deblurring tasks is limited without explicitly modeling blurring as degradation representations. In this paper, we propose IDR2ENet, which is the Implicit Degradation Representations and Reblur Estimation Network, for real image deblurring. IDR2ENet consists of a degradation estimation process, a reblurring process, and a deblurring process. The degradation estimation process takes the real blurred image as input and outputs the implicit degradation representations estimated on it, which are used as the inputs of both reblurring and deblurring processes to better estimate the features of the blurred image. The experimental results show that whether compared with traditional or deep-learning-based deblurring algorithms, IDR2ENet achieves stable and efficient deblurring results on real blurred images.

Medical Image Denoising with Wiener Filter and High Boost Filtering

The Wiener filter is widely used in image de-noising. It is used to reduce Gaussian noise. Although the Wiener filter removes noise from the image, it causes a loss of edge detail information, resulting in blurring of the image. The edge details are considered high-frequency components. The Wiener filter is unable to reconstruct these components. In this paper, the proposed filter based on the Wiener filter and the high-boost filter for medical images is presented. The proposed filter is applied to the degraded image. First, using Fourier Transformation, the degraded image and the high boost filter are converted in the frequency domain. Secondly, the wiener filter is applied to the image along with the high boost filter. Thirdly, the deconvolution process is achieved on the image with the high boost filter. Finally, to reconstruct the sharper image in the spatial domain, the inverse of the Fourier transformation is applied. The proposed filter works to suppress the additive noise at the same time. It can keep the image's edge details. Some focus operators are used, which are image contrast, gradient energy, histogram entropy, and spatial frequency, in order to test the proposed algorithm. Experimental results showed that the proposed filter gives good results compared with the traditional filters for medical images, especially dark images.

Feasibility and clinical usefulness of deep learning-accelerated MRI for acute painful fracture patients wearing a splint: A prospective comparative study.

To evaluate the feasibility and clinical usefulness of deep learning (DL)-accelerated turbo spin echo (TSEDL) sequences relative to standard TSE sequences (TSES) for acute radius fracture patients wearing a splint. This prospective consecutive study investigated 50 patients' preoperative wrist MRI scans acquired between July 2021 and January 2022. Examinations were performed at 3 Tesla MRI with body array coils due to the wrist splint. Besides TSES obtained according to the routine protocol, TSEDL sequences for axial T2-, coronal T1-, and coronal PD-weighted TSE sequences were scanned for comparison. For quantitative assessment, the relative signal-to-noise ratio (rSNR), the relative contrast-to-noise ratio (rCNR), and the relative contrast ratio (rCR) were measured. For qualitative assessment, all images were assessed by two independent musculoskeletal radiologists in terms of perceived SNR, image contrast, image sharpness, artifacts disturbing evaluation, overall image quality and diagnostic confidence for injuries using a four- or five-point Likert scale. The scan time was shortened approximately by a factor of two for TSEDL compared to TSES. TSEDL images showed significantly better rSNR, rCNR, and rCR values for all sequences, and scored significantly better in terms of both image quality and diagnostic confidence for both readers than TSES images (all p < .05). Interrater reliabilities were in almost perfect agreement. The DL-accelerated technique proved to be very helpful not only to reduce scan time but also to improve image quality for acute painful fracture patients wearing a splint despite using body array coils instead of a wrist-specific coil. Based on our study, the DL-accelerated technique can be very useful for MRI of any part of the extremities in trauma settings just with body array coils.

Intriguing Findings of Frequency Selection for Image Deblurring

Blur was naturally analyzed in the frequency domain, by estimating the latent sharp image and the blur kernel given a blurry image. Recent progress on image deblurring always designs end-to-end architectures and aims at learning the difference between blurry and sharp image pairs from pixel-level, which inevitably overlooks the importance of blur kernels. This paper reveals an intriguing phenomenon that simply applying ReLU operation on the frequency domain of a blur image followed by inverse Fourier transform, i.e., frequency selection, provides faithful information about the blur pattern (e.g., the blur direction and blur level, implicitly shows the kernel pattern). Based on this observation, we attempt to leverage kernel-level information for image deblurring networks by inserting Fourier transform, ReLU operation, and inverse Fourier transform to the standard ResBlock. 1 × 1 convolution is further added to let the network modulate flexible thresholds for frequency selection. We term our newly built block as Res FFT-ReLU Block, which takes advantages of both kernel-level and pixel-level features via learning frequency-spatial dual-domain representations. Extensive experiments are conducted to acquire a thorough analysis on the insights of the method. Moreover, after plugging the proposed block into NAFNet, we can achieve 33.85 dB in PSNR on GoPro dataset. Our method noticeably improves backbone architectures without introducing many parameters, while maintaining low computational complexity. Code is available at https://github.com/DeepMed-Lab/DeepRFT-AAAI2023.

Deep Learning-reconstructed Parallel Accelerated Imaging for Knee MRI

Deep learning (DL) can improve image quality by removing noise from accelerated MRI. To compare the quality of various accelerated imaging applications in knee MRI with and without DL. We analyzed 44 knee MRI scans from 38 adult patients using the DL-reconstructed parallel acquisition technique (PAT) between May 2021 and April 2022. The participants underwent sagittal fat-saturated T2-weighted turbo-spin-echo accelerated imaging without DL (PAT-2 [2-fold parallel accelerated imaging], PAT-3, and PAT-4) and with DL (DL with PAT-3 [PAT-3DL] and PAT-4 [PAT-4DL]). Two readers independently evaluated subjective image quality (diagnostic confidence of knee joint abnormalities, subjective noise and sharpness, and overall image quality) using a 4-point grading system (1-4, 4=best). Objective image quality was assessed based on noise (noise power) and sharpness (edge rise distance). The mean acquisition times for PAT-2, PAT-3, PAT-4, PAT-3DL, and PAT-4DL sequences were 2:55, 2:04, 1:33, 2:04, and 1:33 min, respectively. Regarding subjective image quality, PAT-3DL and PAT-4DL scored higher than PAT-2. Objectively, DL-reconstructed imaging had significantly lower noise than PAT-3 and PAT-4 (P <0.001), but the results were not significantly different from those for PAT-2 (P >0.988). Objective image sharpness did not differ significantly among the imaging combinations (P =0.470). The inter-reader reliability ranged from good to excellent (κ = 0.761–0.832). PAT-4DL imaging in knee MRI exhibits similar subjective image quality, objective noise, and sharpness levels compared with conventional PAT-2 imaging, with an acquisition time reduction of 47%.

Dual-Domain Attention for Image Deblurring

As a long-standing and challenging task, image deblurring aims to reconstruct the latent sharp image from its degraded counterpart. In this study, to bridge the gaps between degraded/sharp image pairs in the spatial and frequency domains simultaneously, we develop the dual-domain attention mechanism for image deblurring. Self-attention is widely used in vision tasks, however, due to the quadratic complexity, it is not applicable to image deblurring with high-resolution images. To alleviate this issue, we propose a novel spatial attention module by implementing self-attention in the style of dynamic group convolution for integrating information from the local region, enhancing the representation learning capability and reducing computational burden. Regarding frequency domain learning, many frequency-based deblurring approaches either treat the spectrum as a whole or decompose frequency components in a complicated manner. In this work, we devise a frequency attention module to compactly decouple the spectrum into distinct frequency parts and accentuate the informative part with extremely lightweight learnable parameters. Finally, we incorporate attention modules into a U-shaped network. Extensive comparisons with prior arts on the common benchmarks show that our model, named Dual-domain Attention Network (DDANet), obtains comparable results with a significantly improved inference speed.

Large field-of-view holographic Maxwellian display based on spherical crown diffraction.

Maxwellian display, as an effective solution to the vergence accommodation conflict in near-eye displays (NEDs), has demonstrated its unique advantages in many aspects, such as the ability to provide sharp images within a certain depth of field (DOF) without being affected by the eye's focus. In recent years, the appearance of holographic Maxwellian displays has addressed the shortcomings of traditional Maxwellian displays, meeting the demands for flexible control parameters, aberration-free designing, and expanded eyebox. Nonetheless, the human eye's requirement for immersion still leaves room for a significant improvement in terms of the field-of-view (FOV). In this paper, we propose a large FOV holographic Maxwellian display based on spherical crown diffraction. The proposed spherical-crown holographic Maxwellian display theoretically can cover the full FOV required by the human eyes without complex optical paths and has flexible control of performance parameters such as DOF and image quality. We have successfully demonstrated the feasibility of the spherical crown diffraction model in lensless holographic Maxwellian displays, and it is expected to have practical applications in the field of holographic Maxwellian NEDs in the future.

A Feasibility Study on Deep Learning Reconstruction to Improve Image Quality With PROPELLER Acquisition in the Setting of T2-Weighted Gynecologic Pelvic Magnetic Resonance Imaging.

Evaluate deep learning (DL) to improve the image quality of the PROPELLER (Periodically Rotated Overlapping Parallel Lines with Enhanced Reconstruction technique) for 3 T magnetic resonance imaging of the female pelvis. Three radiologists prospectively and independently compared non-DL and DL PROPELLER sequences from 20 patients with a history of gynecologic malignancy. Sequences with different noise reduction factors (DL 25%, DL 50%, and DL 75%) were blindly reviewed and scored based on artifacts, noise, relative sharpness, and overall image quality. The generalized estimating equation method was used to assess the effect of methods on the Likert scales. Quantitatively, the contrast-to-noise ratio and signal-to-noise ratio (SNR) of the iliac muscle were calculated, and pairwise comparisons were performed based on a linear mixed model. P values were adjusted using the Dunnett method. Interobserver agreement was assessed using the κ statistic. P value was considered statistically significant at less than 0.05. Qualitatively, DL 50 and DL 75 were ranked as the best sequences in 86% of cases. Images generated by the DL method were significantly better than non-DL images (P < 0.0001). Iliacus muscle SNR on DL 50 and DL 75 was significantly better than non-DL images (P < 0.0001). There was no difference in contrast-to-noise ratio between the DL and non-DL techniques in the iliac muscle. There was a high percent agreement (97.1%) in terms of DL sequences' superior image quality (97.1%) and sharpness (100%) relative to non-DL images. The utilization of DL reconstruction improves the image quality of PROPELLER sequences with improved SNR quantitatively.