De-noising Effect Research Articles

Low Dose CT Image Denoising: A Comparative Study of Deep Learning Models and Training Strategies

Article Low Dose CT Image Denoising: A Comparative Study of Deep Learning Models and Training Strategies Heng Zhao 1, Like Qian 1, Yaqi Zhu 1 and Dingcheng Tian 1,2,∗ 1 Research Institute for Medical and Biological Engineering, Ningbo University, Ningbo 315211, China 2 College of Medicine and Biological Information Engineering, Northeastern University, Shenyang 110016, China ∗ Correspondence: 2310520@stu.neu.edu.cn Received: 8 August 2024; Revised: 10 October 2024; Accepted: 14 October 2024; Published: 5 November 2024 Abstract: Low-dose computed tomography (LDCT) denoising is an important topic in CT image research. Compared with normal-dose CT images, LDCT can reduce the radiation dose of X-rays, decreasing the radiation burden on the human body, which is beneficial to human health. However, quantum noise caused by low-dose rays will reduce the quality of CT images, thereby decreasing the accuracy of clinical diagnosis. In recent years, deep learning-based denoising methods have shown promising advantages in this field. Researchers have proposed some optimized models for low-dose CT image denoising. These methods have enhanced the application of low-dose CT image denoising from different aspects. From the perspective of experimental research, this paper investigates and evaluates some top deep learning models proposed in the field of low-dose image denoising in recent years, with the aim of determining the best models and training strategies for this task. We conducted experiments on seven deep learning models (REDCNN, EDCNN, QAE, OCTNet, UNet, WGAN, CTformer) on the AAPM dataset and the Piglet dataset. Our research shows that UNet has the best denoising effect among the models, obtaining PSNR = 33.06 (AAPM dataset) and PSNR = 31.21 (Piglet dataset), and good generalization capacity is also observed. However, UNet has a large number of parameters, and the time it takes to process an image is about 8 ms, while EDCNN takes about 4.8 ms to process an image, and its average PSNR is ranked second after UNet. EDCNN strikes a balance between denoising performance and processing efficiency, making it ideal for low-dose CT image denoising tasks.

A hypothetical defenses-based training framework for generating transferable adversarial examples

Transfer-based attacks utilize the proxy model to craft adversarial examples against the target model and make significant advancements in the realm of black-box attacks. Recent research suggests that these attacks can be enhanced by incorporating adversarial defenses into the training process of adversarial examples. Specifically, adversarial defenses supervise the training process, forcing the attacker to overcome greater challenges and produce more robust adversarial examples with enhanced transferability. However, current methods mainly rely on limited input transformation defenses, which apply only linear affine changes. These defenses are insufficient for effectively removing harmful content from adversarial examples, resulting in restricted improvements in their transferability. To address this issue, we propose a novel training framework named Transfer-based Attacks through Hypothesis Defense (TA-HD). This framework enhances the generalization of adversarial examples by integrating a hypothesis defense mechanism into the proxy model. Specifically, we propose an input denoising network as the hypothesis defense to effectively remove harmful noise from adversarial examples. Furthermore, we introduce an adversarial training strategy and design specific adversarial loss functions to optimize the input denoising network’s parameters. The visualization of the training process demonstrates the effective denoising capability of the hypothesized defense mechanism and the stability of the training process. Extensive experiments show that the proposed training framework significantly improves the success rate of transfer-based attacks by up to 19.9%. The code is available at https://github.com/haolingguang/TA-HD.

Self-supervised learning for effective denoising of flow fields

In this study, we proposed an efficient approach based on a deep learning (DL) denoising autoencoder (DAE) model for denoising noisy flow fields. The DAE operates on a self-learning principle and does not require clean data as training labels. Furthermore, investigations into the denoising mechanism of the DAE revealed that its bottleneck structure with a compact latent space enhances denoising efficacy. Meanwhile, we also developed a deep multiscale DAE for denoising turbulent flow fields. Furthermore, we used conventional noise filters to denoise the flow fields and performed a comparative analysis with the results from the DL method. The effectiveness of the proposed DL models was evaluated using direct numerical simulation data of laminar flow around a square cylinder and turbulent channel flow data at various Reynolds numbers. For every case, synthetic noise was augmented in the data. A separate experiment used particle-image velocimetry data of laminar flow around a square cylinder containing real noise to test DAE denoising performance. Instantaneous contours and flow statistical results were used to verify the alignment between the denoised data and ground truth. The findings confirmed that the proposed method could effectively denoise noisy flow data, including turbulent flow scenarios. Furthermore, the proposed method exhibited excellent generalization, efficiently denoising noise with various types and intensities.

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

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

Denoising of Photon-Counting LiDAR Bathymetry Based on Adaptive Variable OPTICS Model and Its Accuracy Assessment

Spaceborne photon-counting LiDAR holds significant potential for shallow-water bathymetry. However, the received photon data often contain substantial noise, complicating the extraction of elevation information. Currently, a denoising algorithm named ordering points to identify the clustering structure (OPTICS) draws people’s attention because of its strong performance under high background noise. However, this algorithm’s fixed input variables can lead to inaccurate photon distribution parameters in areas near the water bottom, which results in inadequate denoising in these areas, affecting bathymetric accuracy. To address this issue, an Adaptive Variable OPTICS (AV-OPTICS) model is proposed in this paper. Unlike the traditional OPTICS model with fixed input variables, the proposed model dynamically adjusts input variables based on point cloud distribution. This adjustment ensures accurate measurement of photon distribution parameters near the water bottom, thereby enhancing denoising effects in these areas and improving bathymetric accuracy. The findings indicate that, compared to traditional OPTICS methods, AV-OPTICS achieves higher F1-values and lower cohesions, demonstrating better denoising performance near the water bottom. Furthermore, this method achieves an average MAE of 0.28 m and RMSE of 0.31 m, indicating better bathymetric accuracy than traditional OPTICS methods. This study provides a promising solution for shallow-water bathymetry based on photon-counting LiDAR data.

Agent-SwinPyramidNet: an enhanced deep learning model with AMTCF-VMD for anomaly detection in oil and gas pipelines

PurposeThe anomaly detection task for oil and gas pipelines based on acoustic signals faces issues such as background noise coverage, lack of effective features, and small sample sizes, resulting in low fault identification accuracy and slow efficiency. The purpose of this paper is to study an accurate and efficient method of pipeline anomaly detection.Design/methodology/approachFirst, to address the impact of background noise on the accuracy of anomaly signals, the adaptive multi-threshold center frequency variational mode decomposition method(AMTCF-VMD) method is used to eliminate strong noise in pipeline signals. Secondly, to address the strong data dependency and loss of local features in the Swin Transformer network, a Hybrid Pyramid ConvNet network with an Agent Attention mechanism is proposed. This compensates for the limitations of CNN’s receptive field and enhances the Swin Transformer’s global contextual feature representation capabilities. Thirdly, to address the sparsity and imbalance of anomaly samples, the SpecAugment and Scaper methods are integrated to enhance the model’s generalization ability.FindingsIn the pipeline anomaly audio and environmental datasets such as ESC-50, the AMTCF-VMD method shows more significant denoising effects compared to wavelet packet decomposition and EMD methods. Additionally, the model achieved 98.7% accuracy on the preprocessed anomaly audio dataset and 99.0% on the ESC-50 dataset.Originality/valueThis paper innovatively proposes and combines the AMTCF-VMD preprocessing method with the Agent-SwinPyramidNet model, addressing noise interference and low accuracy issues in pipeline anomaly detection, and providing strong support for oil and gas pipeline anomaly recognition tasks in high-noise environments.

Collaborative masking based speckle disentanglement for self-supervised optical coherence tomography image despeckling

Optical coherence tomography (OCT) has proven to be an effective and safe diagnostic tool in clinical settings due to its unique advantages. Nonetheless, the OCT images are unavoidably corrupted with speckle noise. Despeckling has become a vital preliminary step in OCT based image processing and analysis tasks. The supervised deep learning (DL) based denoising algorithms have recently attracted much attention due to their promising performance, but the lack of the noise-free OCT images renders it difficult to provide the effective supervision for DL model training. This paper has presented a self-supervised OCT image despeckling approach which does not depend on the reference clean image and combines the disentangled representation strategy with speckle noise estimation for the effective denoising. The proposed algorithm firstly disentangles the corrupted image into the speckle noise and clean image. Then, the multiple generated clean images are multiplied by the disentangled speckle noise to synthesize the speckle-corrupted image for self-supervised learning. To make the generated clean image and extracted speckle noise more statistically realistic, a collaborative masking strategy and a noise calibration module are applied to explore the spatial correlation of the clean image and refine speckle noise. The proposed method has been validated on two public datasets. Experimental results demonstrate that our method has superior speckle reduction performance to some traditional and DL-based despeckling approaches in terms of visual evaluation and quantitative metrics. Particularly, our method provides contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR) improvements of at least 0.35 dB and 14.37 dB over the compared methods, respectively. Additionally, the retinal image classification results provide further evidence for the exceptional despeckling performance of our approach against other compared methods. It is obvious that the developed method has significant advantages in keeping the fidelity of retinal structure and enhancing the speckle suppression level. Meanwhile, the presented method holds certain reference value for improving diagnostic efficiency in clinical.

Fast audio signal denoising with low-rank approximation using windowed singular value decomposition of frequency division

The method of constructing Hankel matrices and using low-rank approximation has been demonstrated to be an effective approach to audio denoising. However, the significant computational complexity and the trade-off between the denoising quality and the loss of effective signal remain open issues. This paper proposes a denoising method based on frequency-divided Windowed Singular Value Decomposition (WSVD) and exploits the low-rank characteristics and frequency features commonly found in audio signals including speech and music recordings. The method incorporates an improved Lanczos Bidiagonalization algorithm to accelerate the singular value decomposition with low error and high tolerance. Furthermore, techniques are added at the window junctions to maintain the continuity and smoothness of the final audio, thus achieving denoising efficiently and effectively. This paper also assesses the influence of window segmentation length, main frequency domain characteristics, rank selection of Hankel matrix and characteristics of different noises on the final denoising effect. Finally, the denoising algorithm's robustness and effectiveness are validated through simulations and experiments.

DeCoGAN: MVCT image denoising via coupled generative adversarial network

Objective. In helical tomotherapy, image-guided radiotherapy employs megavoltage computed tomography (MVCT) for precise targeting. However, the high voltage of megavoltage radiation introduces substantial noise, significantly compromising MVCT image clarity. This study aims to enhance MVCT image quality using a deep learning-based denoising method. Approach. We propose an unpaired MVCT denoising network using a coupled generative adversarial network framework (DeCoGAN). Our approach assumes that a universal latent code within a shared latent space can reconstruct any given pair of images. By employing an encoder, we enforce this shared-latent space constraint, facilitating the conversion of low-quality (noisy) MVCT images into high-quality (denoised) counterparts. The network learns the joint distribution of images from both domains by leveraging samples from their respective marginal distributions, enhanced by adversarial training for effective denoising. Main Results. Compared to an analytical algorithm (BM3D) and three deep learning-based methods (RED-CNN, WGAN-VGG and CycleGAN), the proposed method excels in preserving image details and enhancing human visual perception by removing most noise and retaining structural features. Quantitative analysis demonstrates that our method achieves the highest peak signal-to-noise ratio and Structural Similarity Index Measurement values, indicating superior denoising performance. Significance. The proposed DeCoGAN method shows remarkable MVCT denoising performance, making it a promising tool in the field of radiation therapy.

Innovative design of wood texture images for indoor furniture based on variable space

In the design of furniture wood texture images, image restoration is a key issue. This study proposes a Bregmanized operator splitting optimization algorithm based on variable space. This study combines variable spatial morphology to process texture images and effectively extract image features using different operators, thereby achieving image restoration. The results of comparing the proposed algorithm with other image processing algorithms showed that the research algorithm achieved a peak signal-to-noise ratio of 29.86 and a structural similarity index of 0.87 in image denoising, respectively, and had a good denoising effect. In terms of image deblurring, the research algorithm had the lowest root mean square error values on the France and Boat datasets, with values of 8.98 and 8.82, respectively, indicating that the image processed by the algorithm had a high similarity with the real image. In terms of image resolution reconstruction, the peak signal-to-noise ratio and root mean square error values of the research algorithm reached 29.74 and 12.67, respectively, indicating that the reconstructed image had the best fit with the original image and the smallest error. In summary, the proposed algorithm has shown good performance in image processing and can be effectively applied in fields such as image denoising, deblurring, and resolution reconstruction. It provides effective methods and technical support for innovative design of wood texture images in indoor furniture.

FPGA-BASED IMPLEMENTATION OF A GAUSSIAN SMOOTHING FILTER WITH POWERS-OF-TWO COEFFICIENTS

The purpose of the study is to develop methods for synthesizing a Gaussian filter that ensures simplified hardware and software implementation, particularly filters with powers-of-two coefficients. Such filters can provide effective denoising of images, including landscape maps, both natural and synthetically generated. The study also involves analyzing of methods for FPGA implementation, comparing their hardware complexity, performance, and noise reduction with traditional Gaussian filters. Results. An algorithm for rounding filter coefficients to powers of two, providing optimal approximation of the constructed filter to the original, is presented, along with examples of developed filters. Topics covered include FPGA implementation, based on the Xilinx Artix-7 FPGA. Filter structures, testing methods, simulation results, and verification of the scheme are discussed. Examples of the technological placement of the implemented scheme on the FPGA chip are provided. Comparative evaluations of FPGA resources and performance for proposed and traditional Gaussian filters are carried out. Digital modeling of the filters and noise reduction estimates for noisy images of the terrain surface are presented. The developed algorithm provides approximation of Gaussian filter coefficients as powers of two for a given window size and maximum number of bits with a relative error of no more than 0.18. Implementing the proposed filters on FPGA results in a hardware costs reduction with comparable performance. Computer simulation show that Gaussian filters both traditional and proposed effectively suppress additive white noise in images. Proposed filters improve the signal-to-noise ratio within 5-10 dB and practically match the filtering quality of traditional Gaussian filters.

Application and characterization of digital Sallen–Key filter in nuclear spectrum smoothing

In the nuclear spectrum analysis processing, spectrum smoothing can remove the statistical fluctuation in the spectrum, which is beneficial for peak detection and peak area calculation. In this work, a spectrum smoothing algorithm is proposed based on digital Sallen–Key filter, which contains four parameters (m, n, k, D). The amplitude–frequency response curve of Sallen–Key filter is deduced and the filtering performance is analyzed. Meanwhile, the effects of the four parameters on the shape of the smoothed spectrum are explored: D affects the counts and peak areas of the spectrum, and the peak area can be corrected by the peak area correction function S’. The parameters of m, n and k affect the peak position after smoothing, making the peak position shift to the right, and the peak position correction function P’ can be used to correct the peak position, when n¿2, the spectrum data appear negative after smoothing, when k¿2, the smoothed spectrum broadening degree is greater than 20%. Smoothness (R), noise smoothing factor (NSF), spectrum count ratio before and after smoothing (PER), and comprehensive evaluation factor (Q) are used to evaluate the smoothing effect of the algorithm. The parameters of the algorithm are optimally selected: about the gamma spectrum of 137Cs and 60Co, the optimal parameters are m=1.5 n=2 k=2 D=1, about the characteristic X-ray spectrum of Fe and quasi-geological sample (TiMnFeNiCuZn), the optimal parameters are m=1.1 n=1.1 k=1.3 D=1. Based on Sallen–Key smoothing method, Fourier transform method, Gaussian function method, wavelet transformation method, center of gravity method and least squares method, the gamma spectrum of 137Cs is smoothed and denoised in this paper. The results show that the Sallen–Key method has better spectrum denoising effect (R=0.6056) and comprehensive performance indicators (Q=0.6104), which can be further applied for the smoothing of nuclear spectrum data.

An intelligent wireless channel corrupted image-denoising framework using symmetric convolution-based heuristic assisted residual attention network

ABSTRACT Image denoising is one of the significant approaches for extracting valuable information in the required images without any errors. During the process of image transmission in the wireless medium, a wide variety of noise is presented to affect the image quality. For efficient analysis, an effective denoising approach is needed to enhance the quality of the images. The main scope of this research paper is to correct errors and remove the effects of channel degradation. A corrupted image denoising approach is developed in wireless channels to eliminate the bugs. The required images are gathered from wireless channels at the receiver end. Initially, the collected images are decomposed into several regions using Adaptive Lifting Wavelet Transform (ALWT) and then the “Symmetric Convolution-based Residual Attention Network (SC-RAN)” is employed, where the residual images are obtained by separating the clean image from the noisy images. The parameters present are optimized using Hybrid Energy Golden Tortoise Beetle Optimizer (HEGTBO) to maximize efficiency. The image denoising is performed over the obtained residual images and noisy images to get the final denoised images. The numerical findings of the developed model attain 31.69% regarding PSNR metrics. Thus, the analysis of the developed model shows significant improvement.

Value of vendor-agnostic deep learning image denoising in brain computed tomography: A multi-scanner study.

To evaluate the effect of a vendor-agnostic deep learning denoising (DLD) algorithm on diagnostic image quality of non-contrast cranial computed tomography (ncCT) across five CT scanners.This retrospective single-center study included ncCT data of 150 consecutive patients (30 for each of the five scanners) who had undergone routine imaging after minor head trauma. The images were reconstructed using filtered back projection (FBP) and a vendor-agnostic DLD method. Using a 4-point Likert scale, three readers performed a subjective evaluation assessing the following quality criteria: overall diagnostic image quality, image noise, gray matter-white matter differentiation (GM-WM), artifacts, sharpness, and diagnostic confidence. Objective analysis included evaluation of noise, contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR), and an artifact index for the posterior fossa.In subjective image quality assessment, DLD showed constantly superior results compared to FBP in all categories and for all scanners (p<0.05) across all readers. The objective image quality analysis showed significant improvement in noise, SNR, and CNR as well as for the artifact index using DLD for all scanners (p<0.001).The vendor-agnostic deep learning denoising algorithm provided significantly superior results in the subjective as well as in the objective analysis of ncCT images of patients with minor head trauma concerning all parameters compared to the FBP reconstruction. This effect has been observed in all five included scanners. · Significant improvement of image quality for 5 scanners due to the vendor-agnostic DLD. · Subjects were patients with routine imaging after minor head trauma. · Reduction of artifacts in the posterior fossa due to the DLD. · Access to improved image quality even for older scanners from different vendors. · Kapper C, Müller L, Kronfeld A et al. Value of vendor-agnostic deep learning image denoising in brain computed tomography: A multi-scanner study. Fortschr Röntgenstr 2024; DOI 10.1055/a-2290-4781.

WaveletDFDS-Net: A Dual Forward Denoising Stream Network for Low-Dose CT Noise Reduction

The challenge of denoising low-dose computed tomography (CT) has garnered significant research interest due to the detrimental impact of noise on CT image quality, impeding diagnostic accuracy and image-guided therapies. This paper introduces an innovative approach termed the Wavelet Domain Dual Forward Denoising Stream Network (WaveletDFDS-Net) to address this challenge. This method ingeniously combines convolutional neural networks and Transformers to leverage their complementary capabilities in feature extraction. Additionally, it employs a wavelet transform for efficient image downsampling, thereby preserving critical information while reducing computational requirements. Moreover, we have formulated a distinctive dual-domain compound loss function that significantly enhances the restoration of intricate details. The performance of WaveletDFDS-Net is assessed by comparative experiments conducted on public CT datasets, and results demonstrate its enhanced denoising effect with an SSIM of 0.9269, PSNR of 38.1343 and RMSE of 0.0130, superior to existing methods.

Machine Learning-Based Automated Method for Effective De-noising of Magnetocardiography Signals Using Independent Component Analysis

Machine Learning-Based Automated Method for Effective De-noising of Magnetocardiography Signals Using Independent Component Analysis

Intelligent fault diagnosis of rolling bearings under small samples based on lightweight UNet with attention-fused residual block

In rotating machinery, rolling bearings are crucial, and their failure can lead to severe accidents and economic losses. Therefore, fault diagnosis of bearings is necessary to maintain the safe and stable operation of modern machinery and equipment. Currently, data-driven intelligent diagnosis methods are mainstream; however, in practical applications, problems such as insufficient fault samples and strong interference signals often exist. At the same time, a large number of edge-end and mobile devices put higher requirements for the size of the diagnostic model. This study addresses these issues by proposing a lightweight UNet (LWUNet) model, which integrates wavelet packet decomposition (WPD) and attention-fused residual block (AFRB) for fault diagnosis in rolling bearings under varying operating conditions, particularly for small sample sizes. Firstly, WPD is used to decompose and reconstruct the fault signal to achieve effective denoising. Secondly, a LW-UNet is constructed for pixel-level feature learning to reduce the number of parameters and improve the accuracy rate. Thereafter, to further enhance the model feature extraction capability, the AFRB is proposed and embedded into the LWUNet to develop the AFRB-LWUNet model. Finally, the reconstructed signals are input to the proposed model for training, and the model performance is examined using a test set. The proposed method is compared with other fault diagnosis models using small sample data of rolling bearings from the Case Western Reserve University, USA and the University of Paderborn, Germany. The results confirm the higher recognition accuracy, stronger generalization ability, and robustness of the proposed method for small samples under various working conditions and intense noise.

Deep Convolutional Dictionary Learning Denoising Method Based on Distributed Image Patches

To address susceptibility to noise interference in Micro-LED displays, a deep convolutional dictionary learning denoising method based on distributed image patches is proposed in this paper. In the preprocessing stage, the entire image is partitioned into locally consistent image patches, and a dictionary is learned based on the non-local self-similar sparse representation of distributed image patches. Subsequently, a convolutional dictionary learning method is employed for global self-similarity matching. Local constraints and global constraints are combined for effective denoising, and the final denoising optimization algorithm is obtained based on the confidence-weighted fusion technique. The experimental results demonstrate that compared with traditional denoising methods, the proposed denoising method effectively restores fine-edge details and contour information in images. Moreover, it exhibits superior performance in terms of PSNR and SSIM. Particularly noteworthy is its performance on the grayscale dataset Set12. When evaluated with Gaussian noise σ=50, it outperforms DCDicL by 3.87 dB in the PSNR and 0.0012 in SSIM.

Self-supervised dual-domain balanced dropblock-network for low-dose CT denoising

Objective. Self-supervised learning methods have been successfully applied for low-dose computed tomography (LDCT) denoising, with the advantage of not requiring labeled data. Conventional self-supervised methods operate only in the image domain, ignoring valuable priors in the sinogram domain. Recently proposed dual-domain methods address this limitation but encounter issues with blurring artifacts in the reconstructed image due to the inhomogeneous distribution of noise levels in low-dose sinograms. Approach. To tackle this challenge, this paper proposes SDBDNet, an end-to-end dual-domain self-supervised method for LDCT denoising. With the network designed based on the properties of inhomogeneous noise in low-dose sinograms and the principle of moderate sinogram-domain denoising, SDBDNet achieves effective denoising in dual domains without introducing blurring artifacts. Specifically, we split the sinogram into two subsets based on the positions of detector cells to generate paired training data with high similarity and independent noise. These sub-sinograms are then restored to their original size using 1D interpolation and learning-based correction. To achieve adaptive and moderate smoothing in the sinogram domain, we integrate Dropblock, a type of convolution layer with regularization, into SDBDNet, and set a weighted average between the denoised sinograms and their noisy counterparts, leading to a well-balanced dual-domain approach. Main results. Numerical experiments show that our method outperforms popular non-learning and self-supervised learning methods, demonstrating its effectiveness and superior performance. Significance. While introducing a novel high-performance dual-domain self-supervised LDCT denoising method, this paper also emphasizes and verifies the importance of appropriate sinogram-domain denoising in dual-domain methods, which might inspire future work.

Denoising image-based experimental data without clean targets based on deep autoencoders

Linear data-driven methods have demonstrated being effective denoising filters for image-based techniques. On the downside, they are still challenged by conditions of extremely low signal-to-noise ratio. A procedure based on denoising autoencoder (AE) is proposed here. The main challenge is the absence of a “clean” target, thus the training is referred to as “blind”. The method is validated on two synthetic datasets, based on the flow around a circular cylinder, and a fabricated pattern. The reconstruction accuracy is found to be mostly sensitive to the size of the latent space of the AE. A criterion to determine the optimal latent vector size based solely on the information available during the training process is proposed. The reconstruction error is shown to be up to one order of magnitude smaller than the case of a denoising Proper Orthogonal Decomposition. Furthermore, a fast approach using an attention mechanism is proposed. This fast alternative algorithm saves up to 98% training time with only a limited reconstruction accuracy loss if compared to the original approach. Finally, the superiority of the proposed AE blind denoising over linear data-driven filtering has been proved on Pressure Sensitive Paint data of an inclined impinging jet, showing a better performance than the POD denoising.