Conventional Correction Methods Research Articles

An Improved Method Combining CNN and 1D-Var for the Retrieval of Atmospheric Humidity Profiles from FY-4A/GIIRS Hyperspectral Data

FY-4A/GIIRS (Geosynchronous Interferometric Infrared Sounder) is the first infrared hyperspectral atmospheric vertical detector in geostationary orbit. Compared to other similar instruments, it has the advantages of high temporal resolution and stationary relative to the ground. Based on the characteristics of GIIRS observation data, we proposed a humidity profile retrieval method. We fully utilized the information provided by the observation and forecast data, and used the two-dimensional brightness temperature data with the dimension of time and optical spectrum as the input of the CNN (convolution neural network model). Then, the obtained brightness temperature data were shown to be more suitable as the input for the physical retrieval method for humidity than the conventional correction method, improving the accuracy of humidity profile retrieval. We performed two comparative experiments. The first experiment results indicate that, compared to ordinary linear correction and ANN (artificial neural network algorithm) correction, our revised observed brightness temperature data are much closer to the simulated brightness temperature obtained by inputting ERA5 reanalysis data into RTTOV (Radiative Transfer for TOVS). The results of the second experiment indicate that the accuracy of the humidity profile retrieved by our method is higher than that of conventional ANN and 1D-Var (one-dimensional variational algorithm). With ERA5 reanalysis data as the reference value, the RMSE (Root Mean Squared Error) of the humidity profiles by our method is less than 8.2% between 250 and 600 hPa. Our method holds the unique advantage of the high temporal resolution of GIIRS, improves the accuracy of humidity profile retrieval, and proves that the combination of machine learning and the physical method is a compelling idea in the field of satellite atmospheric remote sensing worthy of further exploration.

Highly Accurate, Fully Digital Temperature Sensor With Curvature Correction

In this paper, a highly accurate, fully digital temperature sensor with a curvature correction scheme is proposed. Conventional analog temperature sensors are complex and require a large area. Digital temperature sensors are simple with small area, but they have large inaccuracies due to process variations and curvature errors. In particular, curvature errors become more severe as the technology scales down. Thus, a highly accurate curvature correction method is proposed that can be used even in the latest technology nodes. The proposed curvature correction method achieves lower temperature errors with a smaller area than the conventional curvature correction method by using a simple correction. A temperature error that occurs up to 17 °C before the curvature correction is reduced to a range of −0.6 °C to +0.8 °C after the curvature correction. In addition, the measurement results indicate that the resolution is 0.09 °C /bit in the temperature range from 0 °C to 100 °C. The implementation in 14 nm technology node Xilinx ZCU102 field-programmable gate arrays proves that the proposed temperature sensor is very cost effective, while using only 85 slices. Compared with previous digital temperature sensors, the proposed temperature sensor achieves the smallest error and uses the smallest number of hardware resources while being implemented in a state-of-the-art technology node.

An experimental method to correct low-frequency concentric artifacts in photon counting CT

Large-area photon counting detectors (PCDs) are usually built by tiling multiple semiconductor panels that often have slightly different spectral responses to input x-rays. As a result of this spectral inconsistency, experimental PCD-CT images of large, human-sized objects may show high-frequency ring artifacts and low-frequency band artifacts. Due to the much larger width of the bands compared with the rings, the concentric artifact problem in PCD-CT images of human-sized objects cannot be adequately addressed by conventional CT ring correction methods. This work presents an experimental method to correct the concentric artifacts in PCD-CT. The method is applicable to not only energy-discriminating PCDs with multiple bins but also PCDs with only a single threshold controller. Its principle is similar to the two-step beam hardening correction method, except that the proposed method uses pixel-specific polynomial functions to address the spectral inconsistency problem across the detector plane. The pixel-specific polynomial coefficients were experimentally calibrated using 15 acrylic sheets and 6 aluminum sheets of known thicknesses. The pixel-specific polynomial functions were used to convert the measured PCD-CT projection data to acrylic- and aluminum-equivalent thicknesses that are energy-independent. The proposed method was experimentally evaluated using a human cadaver head and multiple physical phantoms: two of them contain iodine and one phantom contains dual K-edge contrast materials (gadolinium and iodine). The results show that the proposed method can effectively remove the low-frequency concentric artifacts in PCD-CT images while reducing beam hardening artifacts. In contrast, the conventional CT ring correction algorithm did not adequately address the low-frequency band artifacts. Compared with the direct material decomposition-based correction method, the proposed method not only relaxes the requirement of multi-energy bins but also generates images with lower noise and fewer concentric artifacts.

Problems in pore property testing of lignite: Analysis and correction

Due to the low coalification degree of lignite, the colloid-like properties and loose characteristics lead to low energy utilization efficiency and seriously hinder the correct understanding of the original pore structure and its mechanical behavior. As a standard process for physical property test, removing the moisture in the pores by drying will cause the lignite to shrink sharply, destroying its original pore structure. Also, for test methods that need to be performed under high pressure, such as mercury intrusion porosimetry (MIP), the soft and loose texture of lignite will cause it to be significantly compressed, making the test results inaccurate. Aiming at the conventional lignite testing process problems, an optimized scheme suitable for lignite was proposed. By rapid freezing and sublimation at low temperatures, freeze-drying can effectively prevent the collapse of lignite. The solid-state of water protects the lignite's primary structure and shape and minimizes volume shrinkage. The average shrinkage rate of lignite sample after freeze-drying is only 9.72%, while that after hot air-drying is 34.8%. Therefore, dehydration by freeze-drying can be an effective method to protect the pore structure of lignite. Based on correcting the syneresis problem, the piecewise linear interpolation method is used to correct the compression effect of lignite. The corrected result excludes the increased pore volume due to the compression of the lignite skeleton. Also, it avoids the under-correction of the low-pressure part and overcorrection of the high-pressure part in the conventional correction method. Based on the corrected nitrogen adsorption and MIP results, the nuclear magnetic resonance (NMR) surface relaxivity of lignite is calculated. The average surface relaxivity value of the samples is 42.8 μm/s, which is significantly greater than that of medium and high-rank coals. The corrected NMR pore size distribution is closer to the actual pore size distribution of lignite. Finally, based on the analysis and correction of the problems in pore property testing of lignite, the improved workflow suitable for lignite is proposed, which provides some referential significance for the correct understanding of its actual physical property.

A Study on the Retrieval of Temperature and Humidity Profiles Based on FY-3D/HIRAS Infrared Hyperspectral Data

Satellite infrared hyperspectral instruments can obtain a wealth of atmospheric spectrum information. In order to obtain high-precision atmospheric temperature and humidity profiles, we used the traditional One-Dimensional Variational (1D-Var) retrieval algorithm, combined with the information capacity-weight function coverage method to select the spectrum channel. In addition, an Artificial Neural Network (ANN) algorithm was introduced to correct the satellite observation data error and compare it with the conventional error correction method. Finally, to perform the temperature and humidity profile retrieval calculation, we used the FY-3D satellite HIRAS (Hyperspectral Infrared Atmospheric Sounder) infrared hyperspectral data and combined the RTTOV (Radiative Transfer for TOVS) radiative transfer model to build an atmospheric temperature and humidity profile retrieval system. We used data on the European region from July to August 2020 to carry out the training and testing of the retrieval system, respectively, and used the balloon-retrieved sounding data of temperature and humidity published by the University of Wyoming as standard truth values to evaluate the retrieval accuracy. Our preliminary research results show that, compared with the retrieval results of conventional deviation correction, the introduction of ANN algorithm error correction can improve the retrieval accuracy of the retrieval system effectively and the RMSE (Root-Mean-Square Error) of the temperature and humidity has a maximum accuracy of improvement of about 0.5 K (The K represents the thermodynamic temperature unit) and 5%, respectively. The temperature and humidity results obtained by the retrieval system were compared with Global Forecast System (GFS) forecast data. The retrieved temperature RMSE was less than 1.5 K on average, which was better than that for the GFS; the humidity RMSE was less than 15% as a whole, and better than the forecast profile between 100 hpa (1 hpa is 100 pa, the pa represents the air pressure unit) and 600 hpa. Compared with AIRS (Atmospheric Infrared Sounder) products, the result of the retrieval system also had a higher accuracy. The main improvement of the temperature was at 200 hpa and 800 hpa, with maximum accuracy improvements of 2 K and 1.5 K, respectively. The RMSE of the humidity retrieved by the system was also better than the AIRS humidity products at most pressure levels, and the error of maximum difference could reach 15%. After combining the two algorithms, the FY-3D/HIRAS infrared hyperspectral retrieval system could obtain higher-precision temperature and humidity profiles, and relevant results could provide a reference for improving the accuracy of business products.

An Uncertainty-Based Regional Comparative Analysis on the Performance of Different Bias Correction Methods in Statistical Downscaling of Precipitation

Statistical downscaling of General Circulation Models (GCM) simulations is widely used for projecting precipitation at different spatiotemporal scales. However, the downscaling process is linked with different source of uncertainty including structural/parametric uncertainty of the model and output uncertainty. This research proposes a novel framework to assess the parametric uncertainty of downscaling model, and used this framework to assess the performance of different bias correction methods linked to the regression-based statistical downscaling model. The used downscaling framework in the current paper is Statistical Downscaling Model (SDSM). The conventional bias correction method linked with SDSM is the Variance InFlation method (VIF), this paper substitutes this method with three different bias correction methods including Local Intensity Scaling (LOCI), Power Transformation (PT), and Quantile Mapping (QM) to assess the associated parametric and global uncertainty of each method in different climate by using a new approach. The proposed method is applied to six different stations located in Iran and United States with different climate status. Average Relative Interval Length (ARIL), P-level, and Normalized Uncertainty Efficiency (NUE) are used as uncertainty indicators to evaluate the results. Results represent that in every assessed climate class, LOCI, and PT, work better than conventional VIF in both amount and occurrence modules of SDSM framework. More precisely, LOCI works better in station that has wet summer, while PT performs well in the stations where there is no or very limited precipitation in summer. Substituting LOCI with VIF, result in increasing the value of NUE by at least factor of 3 in occurrence and amount model which means the significant reduction in structural uncertainty. Also applying PT in arid regions improves the NUE indicator at least by factor 2 in occurrence and amount model and by factor 3 in output uncertainty assessment, and results in less parametric and output uncertainty. Results illustrate the important role of bias correction approaches in reducing structural, and output uncertainty and improving the statistical efficiency of the downscaling model.

A New Method for Correcting Urbanization-Induced Bias in Surface Air Temperature Observations: Insights From Comparative Site-Relocation Data

The effect of urbanization on surface air temperature (SAT) is one of the most important systematic biases in SAT series of urban stations. Correcting this so-called urbanization bias has the potential to provide accurate basic data for long-term climate change monitoring and research. In the western region of the Yangtze River Delta, 42 meteorological stations with site-relocation history from 2009 to 2018 were selected to analyze the statistical characteristics of the differences in comparative site-relocation daily average SAT. The annual average differences in comparative site-relocation SAT series between the old and the new stations (SATDON) were used to characterize the impact of urbanization bias on the air temperature observation series. Using remote sensing technology, spatial datasets of land-use, landscape, and geometric parameters of the underlying surface in the 5-km buffer zone around the station were established as the observed environmental factors of the site, and the differences in these observed environmental factors (DOEFs) between the old and the new stations were calculated to indicate the change induced by urbanization. Next, multiple linear regression models of SATDON and DOEFs were constructed, showing that the error range of the model for simulated SATDON was 3.66–18.21%, and the average error was 10.09%. Finally, this new correction method (NCM) and conventional correction method (CCM) were applied to the correction of the urbanization bias of SAT series at Hefei station. After comparison, it is found that the NCM could reveal clear contributions of the rapid and slow stages of the urbanization process and resultant environmental changes around the stations to the observed SAT. In summary, the NCM based on remote sensing technology can more reasonably and effectively correct the urbanization bias caused by local human activities, as well as reduce the error caused by the selection of reference stations via the conventional correction method.

A simple and effective method for the accurate extraction of kinetic parameters using differential Tafel plots

The practice of estimating the transfer coefficient (alpha) and the exchange current ({i}_{0}) by arbitrarily placing a straight line on Tafel plots has led to high variance in these parameters between different research groups. Generating Tafel plots by finding kinetic current, {i}_{k} from the conventional mass transfer correction method does not guarantee an accurate estimation of the alpha and {i}_{0}. This is because a substantial difference in values of alpha and {i}_{0} can arise from only minor deviations in the calculated values of {i}_{k}. These minor deviations are often not easy to recognise in polarisation curves and Tafel plots. Recalling the IUPAC definition of alpha , the Tafel plots can be alternatively represented as differential Tafel plots (DTPs) by taking the first order differential of Tafel plots with respect to overpotential. Without further complex processing of the existing raw data, many crucial observations can be made from DTP which is otherwise very difficult to observe from Tafel plots. These for example include a) many perfectly looking experimental linear Tafel plots (R2 > 0.999) can give rise to incorrect kinetic parameters b) substantial differences in values of alpha and {i}_{0} can arise when the limiting current ({i}_{L}) is just off by 5% while performing the mass transfer correction c) irrespective of the magnitude of the double layer charging current ({i}_{mathrm{c}}), the Tafel plots can still get significantly skewed when the ratio of {i}_{0}/{i}_{c} is small. Hence, in order to determine accurate values of alpha and {i}_{0}, we show how the DTP approach can be applied to experimental polarisation curves having well defined {i}_{L}, poorly defined {i}_{L} and no {i}_{L} at all.

Design of a Pseudo-Wide Dynamic Range CMOS Image Sensor by Using the Bidirectional Gamma Curvature Technique

This brief presents a $640\times 480$ CMOS image sensor (CIS) with in-circuit bidirectional gamma correction for use in computer vision applications. The conventional gamma correction method can saturate the bright region in an image, leading to the loss of bright image information. Using the proposed bidirectional gamma correction method, which can be implemented with a log-exponential counter, both bright and dark regions in an image can be simultaneously corrected. The transfer function of an 8-bit single-slope analog-to-digital converter can be selected considering either linear or bidirectional gamma curves with three different slope-change points to enhance the visibility in the pseudo-dynamic range. From the prototype sensor, we observed that the visibility of dark images was increased with proper bidirectional gamma curvatures. The CIS was fabricated in a 110 nm CIS process, and the full chip area was 5900 $\mu \text{m}\,\,\mathbf {\times }\,\,5240\,\,\mu \text{m}$ . The total power consumption was 6 mW at a rate of 15 frames per second with an analog, pixel, and digital supply voltage of 3.3 V, 3.3 V, and 1.5 V, respectively.

Using deep learning for pixel-defect corrections in flat-panel radiography imaging.

Purpose: Flat-panel radiography detectors employ thin-film transistor (TFT) panels to acquire high-quality x-ray images. Pixel defects occur due to circuit shorts or opens in the TFT panel. The defects may degrade the image quality, as well as lower the production yield, and eventually raise the production cost. Hence, it is important to develop an appropriate defect correction algorithm for acquired images. Traditional correction algorithms are based on a complicated adaptive filtering technique, which exploits neighbor pixels, to faithfully preserve the edge components. Because of the complexity of the traditional sophisticated approaches, optimizing their correction performances is difficult. Approach: We considered various pixel-defect correction algorithms based on different deep learning models, such as the artificial neural network (ANN), convolutional neural network (CNN), concatenate CNN, and generative adversarial networks (GAN). We considered two cases of maximal defect sizes, and , and conducted extensive learning experiments to find the best structures of the learning models using the mean square error (MSE) as the loss function. Results: To conduct experiments, practical chest x-ray images were acquired from a general radiography detector. The MSE values of the correction results from ANN, CNN, concatenate CNN, and GAN were 69.40, 75.13, 68.21, and 73.77, respectively, and were much smaller than that of the conventional template match correction method. Conclusions: A concatenate CNN showed the best defect-correction performance. However, ANN could achieve a similar correction performance with much smaller encoding complexity. Therefore, the single-layer ANN can efficiently conduct defect corrections in terms of both correction and complexity.

A new PDAF correction method of CMOS image sensor with Nonacell and Super PD to improve image quality in binning mode

This paper presents a new PDAF correction method to improve the binning mode image quality in the world’s first 0.8um 108 mega pixel CMOS Image Sensor with Samsung Nonacell and Super PD technology. PDAF pixels had been fixed by bad-pixel-correction (BPC), referring to the adjacent non-PDAF pixels in the conventional correction method. We demonstrated a new method, named Dilution mode which output their own seed value within the 3x3 same color-channel pixels to video images and deliver AF information through the separate embedded data. As a result, the PDAF artifact, such as a false color, broken line and dot artifact in a high frequency pattern and overall image detail have dramatically improved in dilution mode.

Comprehensive correction method of airtight coring saturation based on core classification

The indoor comprehensive analysis of core saturation of airtight coring wells is an important part of well logging interpretation. According to the saturation data, the geological reserves can be accurately calculated, and the remaining oil saturation and water-flooded zone in the later stage of the production well can be accurately evaluated. Due to the influence of many factors in the coring process and the experiment process, the sum of the core oil and water saturation is usually not equal to 100%. At present, conventional airtight coring correction method is generally to analyze the oil-water saturation, and then correct the data of the same factors that affect the results. This article combines two methods for saturation correction of XX oilfield in China. For cores with consistent missing factors, mathematical statistics are used to correct the saturation. When most of the rock pores have irreducible water and remaining oil, the phase percolation split method is used for the correction after the experimental analysis. By comparing with the logging interpretation results and the results of adjacent wells, the feasibility of the comprehensive correction method can be verified.

Fast Commutation Error Compensation Method of Sensorless Control for MSCMG BLDC Motor With Nonideal Back EMF

As aerospace applications, high reliability and low power consumption operation of the brushless direct current (BLDC) motor for the magnetically suspended control moment gyro (MSCMG) is the primary consideration. To guarantee reliability, the sensorless control without position sensors is adopted. Due to the phase lag caused by hardware and software, the commutation error is inevitable. In order to reduce losses, this article proposes a fast commutation error compensation method to improve commutation accuracy. First, the control topology of the BLDC motor with a buck converter and the commutation point detection circuit are presented. Then, the conducting back electromotive force (EMF) contained in a novel derived dc-link voltage expression is estimated through an extended state observer (ESO). Finally, the commutation error is extracted from the conducting back EMF for compensation. Different from the conventional closed-loop correction method based on the back EMF symmetry, the proposed commutation error compensation method can obtain the commutation error within only one commutation cycle by a continuous estimation with the ESO. The effectiveness of the proposed fast commutation error compensation method was evaluated with an MSCMG system.

Rapid misalignment correction method in reflective fourier ptychographic microscopy for full field of view reconstruction

Fourier ptychographic microscopy (FPM) is a promising phase retrieval algorithm that does not need interferometry. Recently, reflective FPM has been highlighted as a measuring method for surface inspection in the industry, but using it is challenging because of the limited speed of the conventional misalignment correction methods. Here, we propose a new misalignment correction method to overcome this limitation by using additional 4f imaging and modifying the conventional misalignment method. As a result, about 1 mm2 (8095 × 8095) wide field of view image was reconstructed within 980.3 s.

High-Precision Sensorless Optimal Commutation Deviation Correction Strategy of BLDC Motor With Asymmetric Back EMF

Due to the asymmetry of the actual back electromotive force (back EMF), the performance of the conventional commutation deviation correction method degrades in sensorless control of the brushless direct current (BLDC) motor. This article proposes a high-precision sensorless optimal commutation deviation correction strategy, which considers the asymmetry problems of the back EMF. First, this article discusses the optimal commutation issues and the internal power factor angle based commutation deviation correction method. Then, given the issue of the nonsinusoidal phase current integration and the asymmetric back EMF, a dual improved second-order generalized integrator with a positive sequence component extractor (DISOGI-PSCE) is proposed. The DISOGI-PSCE can effectively deal with the dc component and harmonics in flux-linkage and current integration, as well as the asymmetry of the back EMF. Finally, experimental tests are performed on a magnetically suspended control moment gyroscope BLDC motor, which verified the effectiveness of the proposed high-precision sensorless optimal commutation deviation correction method.

Ring-Artifact Correction With Total-Variation Regularization for Material Images in Photon-Counting CT

We propose a ring-artifact correction method with a compressed sensing for material images obtained with a photon-counting computed tomography (CT) system. The ring-artifacts are caused by nonuniformity of detector properties. Conventional ring-artifact correction methods tend to degrade the quality of images. In contrast, compressed sensing methods can correct ring-artifacts with less degradation of the image quality owing to a priori knowledge that ring-artifacts appeared as stripes in sinograms. In this study, we extend the compressed sensing methods for material sinograms obtained with a photon-counting CT system. This is because material sinograms tend to be simpler and sparser, for which a compressed-sensing method can be more effective. We introduced a cost function with a total variation-regularization term and positivity constraint, and optimized it with a prime-dual splitting method. The feasibility of this method was confirmed by simulations and an experiment. In both the simulations and experiment, the proposed method better corrected the ring artifacts than those on attenuation domain and without a priori knowledge. The comparison with previous methods in literature also showed the same results. These results suggest that our method is effective for correcting ring-artifacts in material images.

Integrated Auxiliary Interferometer for Self-Correction of Nonlinear Tuning in Optical Frequency Domain Reflectometry

A new hardware configuration to correct nonlinear tuning affects in optical frequency domain reflectometry (OFDR) is proposed. Unlike conventional correction methods, the new configuration does not require a separate auxiliary interferometer detection, which may eliminate the need for an extra detection device and an acquisition channel. In the proposed method, an intentional beat signal is introduced at the beginning of the OFDR spectrum, which is treated as an auxiliary interferometer to acquire tunable laser phase information for the nonlinear tuning correction. The proposed method may reduce the overall system cost, reduce the data acquisition time and computational load, and make system configuration simpler by eliminating the need for extra components. In this article, the feasibility of the proposed configuration is demonstrated by correcting nonlinear tuning affects using simulations and experiments in a ∼50 m long optical fiber. Distributed strain and temperature sensing are also performed in a 50 m optical fiber with the proposed new configuration.

Interplane bulk motion analysis and removal based on normalized cross-correlation in optical coherence tomography angiography.

Bulk motion seriously degrades the image quality of optical coherence tomography angiography (OCTA). Conventional correction methods focus on in-plane displacement, while the bulk motion component perpendicular to B-scans also introduces noise. This work first presents an evaluation of this component using a specific scan protocol and an approximate expression derived from peak-normalized cross-correlation values, and then quantitatively assesses how interplane bulk motion noise reduce the sensitivity of cross-sectional angiograms. Finally, we developed a repetitive bulk motion correction method based on the estimated displacements and redundant volume scans. The correction does not require registration and angiogram reconstruction of low flow sensitivity frames, and the results of in vivo mice skin OCTA imaging experiments show that the proposed method can effectively reduce bulk motion noise caused by cardiac and respiratory motion and occasional shaking, and improve OCTA image quality, which has practical significance for clinical OCTA diagnosis and analysis.

K-Space deep learning for reference-free EPI ghost correction.

Nyquist ghost artifacts in echo planar imaging (EPI) are originated from phase mismatch between the even and odd echoes. However, conventional correction methods using reference scans often produce erroneous results especially in high-field MRI due to the nonlinear and time-varying local magnetic field changes. Recently, it was shown that the problem of ghost correction can be reformulated as k-space interpolation problem that can be solved using structured low-rank Hankel matrix approaches. Another recent work showed that data driven Hankel matrix decomposition can be reformulated to exhibit similar structures as deep convolutional neural network. By synergistically combining these findings, we propose a k-space deep learning approach that immediately corrects the phase mismatch without a reference scan in both accelerated and non-accelerated EPI acquisitions. To take advantage of the even and odd-phase directional redundancy, the k-space data are divided into 2 channels configured with even and odd phase encodings. The redundancies between coils are also exploited by stacking the multi-coil k-space data into additional input channels. Then, our k-space ghost correction network is trained to learn the interpolation kernel to estimate the missing virtual k-space data. For the accelerated EPI data, the same neural network is trained to directly estimate the interpolation kernels for missing k-space data from both ghost and subsampling. Reconstruction results using 3T and 7T in vivo data showed that the proposed method outperformed the image quality compared to the existing methods, and the computing time is much faster. The proposed k-space deep learning for EPI ghost correction is highly robust and fast, and can be combined with acceleration, so that it can be used as a promising correction tool for high-field MRI without changing the current acquisition protocol.

Model-Based Noise Reduction in Scatter Correction Using a Deep Convolutional Neural Network for Radiography

In radiography, scattered X-rays cause contrast loss in X-ray images, limiting their clinical usefulness, and are often reduced using several scatter correction methods. One difficulty associated with the existing scatter correction methods is noise amplification due to scatter correction. In this study, we investigated a model-based noise reduction method using the U-Net, which is a deep convolutional neural network proposed for image segmentation, to provide a practical solution to the noise amplification problem in conventional scatter correction methods. In this method, the noise properties of an X-ray image after scatter correction are first analyzed using a Poisson-Gaussian mixture model, a trained U-Net model in which the corresponding noise parameters is used predicts image noise, and the scatter-corrected image is finally recovered by subtracting the predicted image noise. We performed a systematic simulation and an experiment to demonstrate its viability and investigated the image characteristics in terms of several image metrics. Our image results showed that the degradation of the image characteristics by scattered X-rays and noise was effectively removed by using the proposed method.