3D Filtering Research Articles

High-Resolution SAR-to-Multispectral Image Translation Based on S2MS-GAN

Synthetic aperture radar (SAR) has been extensively applied in remote sensing applications. Nevertheless, it is a challenge to process and interpret SAR images. The key to interpreting SAR images lies in transforming them into other forms of remote sensing images to extract valuable hidden remote sensing information. Currently, the conversion of SAR images to optical images produces low-quality results and incomplete spectral information. To address these problems, an end-to-end network model, S2MS-GAN, is proposed for converting SAR images into multispectral images. In this process, to tackle the issues of noise and image generation quality, a TV-BM3D module is introduced into the generator model. Through TV regularization, block-matching, and 3D filtering, these two modules can preserve the edges and reduce the speckle noise in SAR images. In addition, spectral attention is added to improve the spectral features of the generated MS images. Furthermore, we construct a very high-resolution SAR-to-MS image dataset, S2MS-HR, with a spatial resolution of 0.3 m, which is currently the most comprehensive dataset available for high-resolution SAR-to-MS image interpretation. Finally, a series of experiments are conducted on the relevant dataset. Both quantitative and qualitative evaluations demonstrate that our method outperforms several state-of-the-art models in translation performance. The solution effectively facilitates high-quality transitions of SAR images across different types.

Synthesis of High-Selectivity Two-Dimensional Filter Banks Using Sigmoidal Function

This paper explores an application of the sigmoidal function—the complementary integral Gaussian error function (erfc(.))—in two-dimensional (2D) uniform and nonuniform filter bank synthesis. The complementary integral Gaussian error function graph represents a smooth low-pass filter magnitude response. A parameter changes the function slope and increases the magnitude response selectivity. The theory is applied to 2D band-pass filter banks. Exact expressions for the magnitude response parameters are determined. As a result, 2D uniform and nonuniform filter banks with very high selectivity and exact shapes are obtained. Three synthesis examples of 2D filter banks with circular and fan-shaped magnitude responses are provided. The theoretical exposition is supplemented with two examples of image analysis using 2D uniform and nonuniform filter banks. A procedure to reduce the computations in image analysis is proposed. A comparison of filter synthesis between Parks–McLellan’s 2D filters and the erfc(.) demonstrates the significantly shorter calculation time of the proposed method.

Planning for complex inferior vena cava filter retrievals: the implementation and effectiveness of 3D printed models

BackgroundInferior vena cava filter (IVC) retrieval is most often routine but can be challenging with high morbidity in complex cases, especially those with an extended dwelling time. While risk of morbidity in complex retrievals is decreased with advanced filter retrieval techniques, deciding when and which to use these requires detailed pre-procedural planning. The purpose of our study was to evaluate patient-specific 3D printed anatomic IVC filter models for aiding complex IVC filter retrievals.MethodsAll IVC filter retrieval patients between June 2021 and September 2022 at one academic medical hospital were prospectively screened. Nine met criteria for complex retrieval, and their CT images were used to 3D print patient-specific IVC and filter models. Models were used in pre-procedural planning and clinical utility was assessed using the Anatomic Model Utility Likert Questionnaire and estimations of the procedural and fluoroscopy time saved.ResultsThe usage of 3D printed models in pre-procedural planning had high clinical utility based on the Likert questionnaire (Anatomic Model Utility Points 366.7 ± 103.1). Using a model significantly increased confidence in planning (p = 0.03) and modified the treatment plan in seven cases. It also led to cost-efficient use of resources in the procedure suite with estimated reduction in procedure and fluoroscopy time of 29.0 [20.3] (p = 0.003) and 10.2 [6.7] (p = 0.002) minutes, respectively.Conclusion3D printed anatomic models for patients who require complex IVC filter retrieval demonstrated Likert-based high clinical utility and led to estimated reductions of procedural and fluoroscopy time.

Early-stage cardiomegaly detection and classification from X-ray images using convolutional neural networks and transfer learning

Cardiomyopathy is a serious condition that can result in heart failure, sudden cardiac death, malignant arrhythmias, and thromboembolism. It is a significant contributor to morbidity and mortality globally. The initial finding of cardiomegaly on radiological imaging may signal a deterioration of a known heart condition, an unknown heart disease, or a heart complication related to another illness. Further cardiological evaluation is needed to confirm the diagnosis and determine appropriate treatment. A chest radiograph (X-ray) is the main imaging method used to identify cardiomegaly when the heart is enlarged. A prompt and accurate diagnosis is essential to help healthcare providers determine the most appropriate treatment options before the condition worsens. This study aims to utilize convolutional neural networks and transfer learning techniques, specifically Inception, DenseNet-169, and ResNet-50, to classify cardiomegaly from chest X-ray images automatically. The utilization of block-matching and 3D filtering (BM3D) techniques aimed at enhancing image edge retention, decreasing noise, and utilizing contrast limited adaptive histogram equalization (CLAHE) to enhance contrast in low-intensity images. Gradient-weighted Class Activation Mapping (GradCAM) was used to visualize the significant activation regions contributing to the model's decision. After evaluating all the models, the ResNet-50 model showed outstanding performance. It achieved perfect accuracy of 100 % in both training, and validation, and an impressive 99.8 % accuracy in testing. Additionally, it displayed complete 100 % precision, recall, and F1-score. These findings demonstrate that ResNet-50 surpassed all other models in the study. As a result, the impressive performance of the ResNet-50 model suggests that it could be a valuable tool in improving the efficiency and accuracy of cardiomyopathy diagnosis, ultimately leading to better patient outcomes.

An adaptive spatiotemporal filter for ultrasound localization microscopy based on density canopy clustering

Ultrasound Localization Microscopy (ULM) facilitates structural and hemodynamic imaging of microvessels with a resolution of tens of micrometers. In ULM, the extraction of effective microbubble signals is crucial for image quality. Singular Value Decomposition (SVD) is currently the most prevalent method for microbubble signal extraction in ULM. Most existing ULM studies employ a fixed SVD filter threshold using empirical values which will lead to imaging quality degradation due to the insufficient separation of blood signals. In this study, we propose an adaptive and non-threshold SVD filter based on canopy-density clustering, termed DCC-SVD. This filter automatically classifies the components of the SVD based on the density of their spatiotemporal features, eliminating the need for parameter selection. In in vitro tube phantom, DCC-SVD demonstrated its ability to adaptive separation of blood and bubble signal at varying microbubble concentrations and flow rates. We compared the proposed DCC-SVD method with the Block-match 3D (BM3D) filter and a classical adaptive method called spatial similarity matrix (SSM), using concentration-variable in vivo rat brain data, as well as open-source rat kidney and mouse tumor datasets. The proposed DCC-SVD improved the global spatial resolution by approximately 4 μm from 30.39 μm to 26.02 μm. It also captured vessel structure absent in images obtained by other methods and yielded a smoother vessel intensity profile, making it a promising spatiotemporal filter for ULM imaging.

3D‐Printed Highly Porous Functional Materials for the Efficient Removal of Adenovirus

AbstractHierarchical porous acrylate‐based materials are highly interesting as 3D filter materials, such as for virus removal from suspensions. Here, the synthesis of highly porous monolithic 3D materials by polymerization‐induced phase separation in liquid crystal display (LCD) based 3D printing is presented for the efficient removal of human adenovirus type 5. The hierarchical porosity can be tuned via the variation of the photocurable resin composition (i.e., inherent porosity) and the computer‐aided design (i.e., “printed” porosity; microchannels). 3D polymer structures with highly intricate geometries and structural features ranging from ≈20 nm up to cm can be achieved, which can be used for effective virus removal in a laboratory‐scale flow‐through approach. Combined focused ion beam/scanning electron microscopy tomography and mercury porosimetry provide detailed information on the inherent pore size, pore size distribution, and pore interconnectivity, which is key for the performance of such functional 3D materials. Polymers with a theoretical void volume of 75% show virus capture with a removal efficiency of ≈70% of the adenovirus. Polymers with the same theoretical void volume and macroscopic design but a more hydrophobic nature captured only ≈33%. An optimized adenovirus retention of 98% is achieved by adjusting the microchannels of the tunable inserts.

Higher order and optimized symmetric 2D FIR filter design and hardware architecture implementation using CSD-CSE

In this paper, an optimized and high-performance Two-Dimensional (2D) Finite Impulse Response (FIR) filter is designed and hardware architecture is implemented for image processing applications. The higher-order circular symmetric 2D FIR filter is designed using a modified McClellan Transformation called a P4 transformation. The perfect circular symmetry in the contour of the 2D FIR filter and less complexity and error are attained by this proposed P4 Transformation. Next, the designed filter coefficients are represented by the Canonical Signed Digit (CSD) number format to attain the multiplier-less design to avoid the power complexity multipliers. Further, to reduce the hardware complexity, the Common Subexpression Elimination (CSE) technique is utilized to reduce the number of adders and hence area and power are decreased. Each sub-filter corresponding to the CSD coefficient rows is realized and integrated as a 2D FIR filter using a Fully Direct-type architecture. This 2D FIR filter architecture is coded by Verilog and synthesized by Cadence tools in a 45 nm CMOS technology library. The Delay, Power Consumption (PC), and Area reports were generated by the Genus synthesis tool and compared with the state-of-the-art works. The PC, area, and delay of the proposed filter architecture are reduced to the minimum of 13.96%, 69.1%, and 1.4%, respectively when compared to the existing filter architectures. The maximum of 10.48 times and a minimum of 1.18 times of Area Delay Product (ADP), and a maximum of 10.69 times and a minimum of 1.063 times of Power Delay Product (PDP) are smaller than the existing 2D filter architectures.

A dimensionality reduction optimization method for thin-walled parts to realize mass center control and lightweight design

Structural lightweight design has attracted growing attention. However, most studies have ignored the impact of lightweight design on the location of mass centres, which is an important factor affecting the dynamic performance of structures. This article proposes a structural optimization method for thin-walled parts to achieve both mass centre control and lightweight design. First, mass centre and volume are taken as the constraints for structural optimization, and a mathematical model for multi-constrained topology optimization is developed. Secondly, a dimensionality reduction method based on a 3D spatial filter is introduced for the optimization of thin-walled parts. The maximum error between the mass centre of the optimized structure and the target mass centre is only 1.03%, which indicates that the method has high accuracy in mass centre control. Numerical analysis shows that adjusting the mass centre can achieve regulation of the moment of inertia and can also suppress flutter.

A denoising method for power equipment images based on block-matching and 3D filtering.

A substation is important equipment of the power system, and there are many power equipment components in the substation. In order to better detect the working status of power equipment components, it is necessary to preprocess these components. In the actual application, the power equipment images may be noisy due to external environmental interference. Therefore, it should denoise these images in order to improve system detection performance. This paper uses the acquired power equipment images and adds noise intensity of 10, 15, 20, 25, and 30, respectively. Then, the Block-Matching and 3D Filtering (BM3D) method is used to denoise these images. BM3D includes three steps such as block combination, collaborative filtering, and integration, which has strong denoising ability. The experimental results show that the proposed method outperforms other methods in terms of denoising visual effects and evaluation indicators. Especially in terms of preserving details and textures of the denoised image, there is a significant advantage in suppressing strong noise. In summary, the proposed method can achieve encouraging denoising results, which is an effective denoising method for power equipment images.

Weakly supervised low-dose computed tomography denoising based on generative adversarial networks.

Low-dose computed tomography (LDCT) is a diagnostic imaging technique designed to minimize radiation exposure to the patient. However, this reduction in radiation may compromise computed tomography (CT) image quality, adversely impacting clinical diagnoses. Various advanced LDCT methods have emerged to mitigate this challenge, relying on well-matched LDCT and normal-dose CT (NDCT) image pairs for training. Nevertheless, these methods often face difficulties in distinguishing image details from nonuniformly distributed noise, limiting their denoising efficacy. Additionally, acquiring suitably paired datasets in the medical domain poses challenges, further constraining their applicability. Hence, the objective of this study was to develop an innovative denoising framework for LDCT images employing unpaired data. In this paper, we propose a LDCT denoising network (DNCNN) that alleviates the need for aligning LDCT and NDCT images. Our approach employs generative adversarial networks (GANs) to learn and model the noise present in LDCT images, establishing a mapping from the pseudo-LDCT to the actual NDCT domain without the need for paired CT images. Within the domain of weakly supervised methods, our proposed model exhibited superior objective metrics on the simulated dataset when compared to CycleGAN and selective kernel-based cycle-consistent GAN (SKFCycleGAN): the peak signal-to-noise ratio (PSNR) was 43.9441, the structural similarity index measure (SSIM) was 0.9660, and the visual information fidelity (VIF) was 0.7707. In the clinical dataset, we conducted a visual effect analysis by observing various tissues through different observation windows. Our proposed method achieved a no-reference structural sharpness (NRSS) value of 0.6171, which was closest to that of the NDCT images (NRSS =0.6049), demonstrating its superiority over other denoising techniques in preserving details, maintaining structural integrity, and enhancing edge contrast. Through extensive experiments on both simulated and clinical datasets, we demonstrated the superior efficacy of our proposed method in terms of denoising quality and quantity. Our method exhibits superiority over both supervised techniques, including block-matching and 3D filtering (BM3D), residual encoder-decoder convolutional neural network (RED-CNN), and Wasserstein generative adversarial network-VGG (WGAN-VGG), and over weakly supervised approaches, including CycleGAN and SKFCycleGAN.

Accurate recognition of light beams carrying orbital angular momentum through scattering media using ghost diffraction

We report a ghost diffraction-based approach to realize accurate recognition of light beams carrying orbital angular momentum (OAM) through dynamic and complex scattering media. A bit sequence is first encoded into an OAM beam, which is sequentially modulated by a series of Hadamard patterns, and then an optical wave propagates through dynamic and complex scattering media. The collected single-pixel light intensities are temporally corrected, and ghost images can be reconstructed by using the principle of ghost diffraction. The reconstructed ghost images are further processed by using block-matching and 3D filtering with image registration, which are then utilized for OAM recognition assisted by the featured normalized cross correlation. Optical experiments are conducted to demonstrate that light beams carrying OAM can be accurately recognized in dynamic and complex scattering environments, and the proposed approach is feasible and effective. The developed ghost diffraction-based approach could open an avenue for various OAM-encoded applications in dynamic and complex scattering environments.

IE-CycleGAN: improved cycle consistent adversarial network for unpaired PET image enhancement.

Technological advances in instruments have greatly promoted the development of positron emission tomography (PET) scanners. State-of-the-art PET scanners such as uEXPLORER can collect PET images of significantly higher quality. However, these scanners are not currently available in most local hospitals due to the high cost of manufacturing and maintenance. Our study aims to convert low-quality PET images acquired by common PET scanners into images of comparable quality to those obtained by state-of-the-art scanners without the need for paired low- and high-quality PET images. In this paper, we proposed an improved CycleGAN (IE-CycleGAN) model for unpaired PET image enhancement. The proposedmethod is based on CycleGAN, and the correlation coefficient loss and patient-specific prior loss were added to constrain the structure of the generated images. Furthermore, we defined a normalX-to-advanced training strategy to enhance the generalization ability of the network. The proposedmethod was validated on unpaired uEXPLORER datasets and Biograph Vision local hospital datasets. For the uEXPLORER dataset, the proposedmethod achieved betterresults than non-local mean filtering (NLM), block-matching and 3D filtering (BM3D), and deep image prior (DIP), which are comparable to Unet (supervised) and CycleGAN (supervised). For the Biograph Vision local hospital datasets, the proposedmethod achieved higher contrast-to-noise ratios(CNR) and tumor-to-background SUVmax ratios (TBR) than NLM, BM3D, and DIP. In addition, the proposedmethod showed higher contrast, SUVmax,and TBR than Unet (supervised) and CycleGAN (supervised) when applied to images from different scanners. The proposed unpaired PET image enhancementmethod outperforms NLM, BM3D, and DIP. Moreover, it performs better than the Unet(supervised) and CycleGAN(supervised) when implemented on local hospital datasets, which demonstrates its excellent generalization ability.

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

PurposeThis 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. MethodsUsing 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. ResultsThe 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). ConclusionsThe 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.

Enabling Low-Dose In Vivo Benchtop X-ray Fluorescence Computed Tomography through Deep-Learning-Based Denoising.

X-ray Fluorescence Computed Tomography (XFCT) is an emerging non-invasive imaging technique providing high-resolution molecular-level data. However, increased sensitivity with current benchtop X-ray sources comes at the cost of high radiation exposure. Artificial Intelligence (AI), particularly deep learning (DL), has revolutionized medical imaging by delivering high-quality images in the presence of noise. In XFCT, traditional methods rely on complex algorithms for background noise reduction, but AI holds promise in addressing high-dose concerns. We present an optimized Swin-Conv-UNet (SCUNet) model for background noise reduction in X-ray fluorescence (XRF) images at low tracer concentrations. Our method's effectiveness is evaluated against higher-dose images, while various denoising techniques exist for X-ray and computed tomography (CT) techniques, only a few address XFCT. The DL model is trained and assessed using augmented data, focusing on background noise reduction. Image quality is measured using peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM), comparing outcomes with 100% X-ray-dose images. Results demonstrate that the proposed algorithm yields high-quality images from low-dose inputs, with maximum PSNR of 39.05 and SSIM of 0.86. The model outperforms block-matching and 3D filtering (BM3D), block-matching and 4D filtering (BM4D), non-local means (NLM), denoising convolutional neural network (DnCNN), and SCUNet in both visual inspection and quantitative analysis, particularly in high-noise scenarios. This indicates the potential of AI, specifically the SCUNet model, in significantly improving XFCT imaging by mitigating the trade-off between sensitivity and radiation exposure.

Application of silicone nanofilaments coatings towards point-of-use drinking water treatment: Assessment on trace Lead ions, dyes and oils removal

A new coating-based approach for water treatment was demonstrated using silicone nanofilaments (SNF) coatings. In view of some existing point-of-use (POU)/point-of-entry (POE) water treatment systems, where membranes or filters are used to build devices that allow individuals to treat water on their own at the point of consumption. The reported coating-based approach offers an advantage as contaminated water was treated upon contact with the functionalized SNF-coated surfaces, without additional pressure or energy requirement. Besides, the coatings took the shapes of the substrates and demonstrated some different formats can be potentially deployed for POU devices. To illustrate the suggested approach, glass substrates of three different shapes (flat microscopic slides, walls of vials, and three-dimensional (3D) filter frits) were employed. Substrates were readily coated with SNF under ambient conditions via gas phase deposition based on the previous developed protocol, subsequently modified with (3-Mercaptopropyl)triethoxysilane (MPTES), and were used to treat 200 ppb of trace lead ions in water. Results suggested that the coatings as a facile mean to render surfaces/containers functional towards trace lead removal, and enhanced performance compared to their non-coated counterparts.In addition, the oil sorption capacities of the coated frits were also evaluated. By rendering the as-prepared frits amphiphilic, they were used to treat trace lead, water-soluble methylene blue, and oil-soluble Oil Red O dyes. Overall, the study demonstrated and evaluated the use of the SNF coatings by treating some model contaminants in different formats than the existing approaches. The ease of preparation and the flexibility of the approach presented provide opportunities for the design of new water treatment systems and more accessible POU/POE devices.

An Intelligent Computer Aided Diagnosis System for Classification of Ovarian Masses using Machine Learning Approach

Ovarian cancer, a difficult and often asymptomatic malignancy, remains a substantial global health concern in women. An ovary is a female reproductive organ, which lies on each side of the uterus and used to store eggs. Computer-aided diagnosis (CAD) is an approach that involves using computer algorithms and machine learning techniques to assist medical professionals in diagnosing ovarian malignancies, benign tumors or Poly-cystic ovaries (PCOS). The need for models that can effectively predict benign ovarian tumors and ovarian cancer has led to the use of machine learning techniques. Our research objective is to propose a machine learning-based system for accurate and early ovarian mass detection utilizing novel annotated ovarian masses. We have used an actual patient database whose input features were extracted from 187 transvaginal ultrasound images from database. The input image is preprocessed using the Block Matching 3D filter. The process involves employing binary and watershed segmentation techniques, followed by the integration of Gabor, Gray-Level Co-Occurrence Matrix (GLCM), Tamura, and edge feature extraction methods. K-Nearest Neighbors (KNN) and Random Forest (RF) are two classifiers used for classification. Based on our results, we are able to demonstrate that binary segmentation with RF classifiers is more accurate (above 86%) than KNN classifiers (under 84%).

Empowering powdered activated carbon (PAC) with 3D printing: Achieving highly efficient and reusable cationic dye removal

Powdered activated carbon (PAC) has been extensively used as an effective adsorbent. Despite its excellent adsorption ability, PAC has drawbacks, including difficulty in filtration and reactivation after use, limitations of mass transfer in deeper areas because of its aggregated powder form, and limited applicability in high-flow systems. To overcome these limitations, we used a three-dimensional (3D) printing system to fabricate PAC into a 3D structure. Spectral and microscopic analyses indicated that PAC was embedded into 3D monolith and exhibited high porosity suitable for facile mass transfer. The designed 3D PAC filter effectively removed 200 ppm-methylene blue (MB) within 8 h and showed an adsorption efficiency of 93.4 ± 0.9%. The adsorption of MB onto the 3D PAC filter was described by the pseudo-first-order kinetic and Freundlich isotherm models. The negatively charged 3D PAC filter might attract the positively charged MB, thus favoring the physical adsorption of MB onto the 3D PAC filter. The adsorption performance of the 3D PAC filter was tested at various pH levels of 4–10 and against MB spiked in seawaters and freshwaters to evaluate its feasibility for use in real environments. Finally, the reproducibility and reusability of the 3D PAC filter were demonstrated through repeated adsorption and desorption processes against MB.

Analytical Design of Gaussian Anisotropic 2D FIR Filters and Their Implementation Using the Block Filtering Approach

This work proposes an analytical design procedure for a particular class of 2D filters, namely anisotropic Gaussian FIR filters. The design is achieved in the frequency domain and starts from a low-pass Gaussian 1D prototype with imposed specifications, whose frequency response is efficiently approximated by a factored trigonometric polynomial using the Chebyshev series. Then, using specific 1D to 2D frequency mappings applied to the prototype, the frequency response for a 2D anisotropic filter with a specified orientation angle is directly derived in two versions, namely with a straight or elliptical shape in the frequency plane. The resulting filters have an accurate shape with low distortion. Several design examples for specified parameters (angle and selectivity) are provided. Then, simulations of directional filtering on various test images are given, which show their capability of extracting oriented lines or other various oriented objects from synthetic or real-life images. Finally, a computationally efficient implementation at the system level is proposed, based on a polyphase decomposition and block-filtering approach, which yields a 2D filter with a high degree of parallelism and low arithmetic complexity.

Turbulence-resistant free space optical communication via chaotic block-matching and 3D filtering.

In this paper, we propose a chaotic block-matching and three-dimensional (C-BM3D) filtering algorithm to remove the noise and enhance the security in the turbulent channel of free space optical (FSO) communication. We experimentally demonstrate the performance of C-BM3D by comparing it with chaotic non-local means filtering (C-NLM), chaotic Gaussian filtering and chaotic Median filtering based on Log-normal and Gamma-Gamma turbulence models. The results show that the peak signal-to-noise ratios (PSNRs) of C-BM3D in the weak turbulence under Log-normal and Gamma-Gamma models are up to 96.2956 and 93.2853, respectively. The C-BM3D also achieves superior image similarity in Log-normal turbulent channel, with its structural similarity index measures (SSIMs) nearly equal to 1. Additionally, the signal-to-noise ratio (SNR) of C-BM3D ranks the highest, and its bit error rate (BER) improves by at least 15 dB compared to that of the other three algorithms. The experimental results indicate that the C-BM3D can be a good candidate for the next generation of FSO communication in security and turbulence resistance.

Subjective and objective image quality of low-dose CT images processed using a self-supervised denoising algorithm.

This study aimed to assess the subjective and objective image quality of low-dose computed tomography (CT) images processed using a self-supervised denoising algorithm with deep learning. We trained the self-supervised denoising model using low-dose CT images of 40 patients and applied this model to CT images of another 30 patients. Image quality, in terms of noise and edge sharpness, was rated on a 5-point scale by two radiologists. The coefficient of variation, contrast-to-noise ratio (CNR), and signal-to-noise ratio (SNR) were calculated. The values for the self-supervised denoising model were compared with those for the original low-dose CT images and CT images processed using other conventional denoising algorithms (non-local means, block-matching and 3D filtering, and total variation minimization-based algorithms). The mean (standard deviation) scores of local and overall noise levels for the self-supervised denoising algorithm were 3.90 (0.40) and 3.93 (0.51), respectively, outperforming the original image and other algorithms. Similarly, the mean scores of local and overall edge sharpness for the self-supervised denoising algorithm were 3.90 (0.40) and 3.75 (0.47), respectively, surpassing the scores of the original image and other algorithms. The CNR and SNR for the self-supervised denoising algorithm were higher than those for the original images but slightly lower than those for the other algorithms. Our findings indicate the potential clinical applicability of the self-supervised denoising algorithm for low-dose CT images in clinical settings.