Super-resolution Convolutional Neural Network Research Articles

Real-Time CNN Training and Compression for Neural-Enhanced Adaptive Live Streaming.

We propose a real-time convolutional neural network (CNN) training and compression method for delivering high-quality live video even in a poor network environment. The server delivers a low-resolution video segment along with the corresponding CNN for super resolution (SR), after which the client applies the CNN to the segment in order to recover high-resolution video frames. To generate a trained CNN corresponding to a video segment in real-time, our method rapidly increases the training accuracy by promoting the overfitting property of the CNN while also using curriculum-based training. In addition, assuming that the pretrained CNN is already downloaded on the client side, we transfer only residual values between the updated and pretrained CNN parameters. These values can be quantized with low bits in real time while minimizing the amount of loss, as the distribution range is significantly narrower than that of the updated CNN. Quantitatively, our neural-enhanced adaptive live streaming pipeline (NEALS) achieves higher SR accuracy and a lower CNN compression loss rate within a constrained training time compared to the state-of-the-art CNN training and compression method. NEALS achieves 15 to 48% higher quality of the user experience compared to state-of-the-art neural-enhanced live streaming systems.

Medical X-ray images enhancement based on super resolution convolution neural network

<p>Pneumonia is a severe lung infection, chest X-ray (CXR) image preferred to find infection. Real images lost its quality, resolution and other feature due to transmission. So good qualitative datasets are very limited. Quality enhancement in medical images is challenging task for researchers. And quality in clinical diagnosis of any disease in deep learning play a very important role. So, this paper presents an aspect with importance of quality in medical images CXR of a particular dataset and how to enhance and create new images with high quality resolution, that is re-used for classification in deep learning. Super resolution convolutional neural netwok (SRCNN) is deep learning based method, which is used for improving resolution in image. Super resolution means low resolution (LR) images from dataset is to be reconstructed or magnified into high resolution (HR). The objective behind this study is to measure the effect of super resolution with quality index, peak signal-to-noise ratio (PSNR), mean squared error (MSE), and structural similarity index measure (SSIM). This experinment performed on 200 images with 10 batches, each batch has 20 images from Kermany dataset, select LR images and converted into HR with SRCNN. Then we find PSNR value of image is increase upto 2 to 5 DB, and MSE of qood quality images is near to zero and MSE decrease up to 20-25, SSIM value have little variation due to same pattern is found in input and output images. Enhancement means highlight or improve the region of interest of pneumonic images. Main goal of this study is to preapare a modified dataset which is further used for classification.</p>

Optimising signal quality for corrosion detection using the innovative Mag_FSRCNN model to minimise hall sensor array in eddy current testing

ABSTRACT The security of aviation hinges significantly on the strength of aircraft structures, with potential risks arising from issues like corrosion impacting the framework. Eddy current testing (ECT) sensors, particularly those utilising array technology, present a promising solution to address this concern. However, challenges persist in improving the clarity of corrosion signals while optimising associated expenses. In response to this, our study introduces the Mag_FSRCNN model, a novel framework that integrates a Fast Super-Resolution Convolutional Neural Network (FSRCNN) module network with a part of Generator module of Generative Adversarial Network to enhance magnetic images obtained by the ECT sensor array. Mag_FSRCNN surpasses traditional methods, showcasing high-resolution capabilities in enhancing ECT signal images. The model demonstrates excellence when evaluated using metrics such as Structural Similarity Index, Root Mean Square Error, and Peak Signal-to-Noise Ratio. These results significantly contribute to the ongoing progress of ECT sensors, facilitating improved resolution and cost-effective application of Hall sensor arrays in the current evolution of ECT sensor technology.

ESM data downscaling: a comparison of super-resolution deep learning models

Climate projections at fine spatial resolutions are required to conduct accurate risk assessment for critical infrastructure and design adaptation planning. Generating these projections using advanced Earth system models (ESM) requires significant computational resources. To address this issue, various statistical downscaling techniques have been introduced to generate fine-resolution data from coarse-resolution simulations. In this study, we evaluate and compare five deep learning-based downscaling techniques, namely, super-resolution convolutional neural networks, fast super-resolution convolutional neural network ESM, efficient sub-pixel convolutional neural network, enhanced deep residual network (EDRN), and super-resolution generative adversarial network (SRGAN). These techniques are applied to a dataset generated by the Energy Exascale Earth System Model (E3SM), focusing on key surface variables such as surface temperature, shortwave heat flux, and longwave heat flux. Models are trained and validated using paired fine-resolution (0.25∘\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{\\circ }$$\\end{document}) and coarse-resolution (1∘\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$^{\\circ }$$\\end{document}) monthly data obtained from a 9-year simulation. Next, blind testing is performed using monthly data obtained from two different years outside of the training and validation set. To evaluate the efficiency of each technique, different statistical metrics are used, including mean squared error (MSE), peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and learned perceptual image patch similarity (LPIPS). The results show that EDRN outperforms other algorithms in terms of PSNR, SSIM, and MSE, but struggles to capture fine-scale features in the data. In contrast, SRGAN, a generative model that uses perceptual loss, excels in capturing fine details at boundaries and internal structures, resulting in lower LPIPS than other methods.

A reconstruction and recovery network-based channel estimation in high-speed railway wireless communications

The integration of high-speed railway communication systems with 5G technology is widely recognized as a significant development. Due to the considerable mobility of trains and the complex nature of the environment, the wireless channel exhibits non-stationary characteristics and fast time-varying characteristics, which presents significant hurdles in terms of channel estimation. In addition, the use of massive MIMO technology in the context of 5G networks also leads to an increase in the complexity of estimation. To address the aforementioned issues, this paper presents a novel approach for channel estimation in high mobility scenarios using a reconstruction and recovery network. In this method, the time-frequency response of the channel is considered as a two-dimensional image. The Fast Super-Resolution Convolution Neural Network (FSRCNN) is used to first reconstruct channel images. Next, the Denoising Convolution Neural Network (DnCNN) is applied to reduce the channel noise and improve the accuracy of channel estimation. Simulation results show that the accuracy of the channel estimation model surpasses that of the standard channel estimation method, while also exhibiting reduced algorithmic complexity.

PRO-SSRGAN: stable super-resolution generative adversarial network based on parameter reconstructive optimization on Gaofen-5 remote-sensing images

ABSTRACT Advances in hardware capabilities and big data technologies over the past decade have enabled the application of deep learning techniques to address the challenge of super-resolution in remote-sensing images. Recently, while deep-learning-based methods have outperformed traditional methods, the abundance of information in remote-sensing images creates an imbalance between performance and computational resource consumption in current deep-learning-based methods. This paper introduces a stable super-resolution algorithm based on parameter reconstructive optimization to address these issues. First, based on the stable super-resolution generative adversarial network (SSRGAN), the algorithm employs a generator with networks using residual connections to reconstruct images with enhanced resolution. Next, it extracts content, adversarial, and regularization losses using the discriminator from a stable super-resolution generative adversarial network, which in turn guides the training of the network. Finally, upon completion of training, structural re-parameterization is conducted to optimize the multi-branch trained network into a plain inference network. This inference network can serve as an individual new generator. The results of qualitative and quantitative experimental comparisons with the models Bicubic, super-resolution convolutional neural network, very deep super-resolution convolutional network, deep recursive residual network, super-resolution generative adversarial network (SRGAN), enhanced SRGAN, and stable SRGAN on the Gaofen-5 AHSI satellite dataset suggest that this algorithm achieves improved evaluation indices with a 4× magnification ratio, reaching a peak signal-to-noise ratio of 30.7207 dB and structural similarity index measure of 0.8178. Compared with the trained but unconverted generator, which can also work independently, implementing re-parameterization results in approximately a 10% reduction in the number of parameters, indicating lower resource consumption, while the reconstruction effect is minimally influenced. Furthermore, the super-resolution results exhibit richer detail, increased contrast, and better scene adaptability.

Ultrasound super resolution imaging for accurate uterus tumor detection and malignancy prediction

The term ‘tumor’ describes an atypical development of cells that forms a mass inside an organ; depending on the organ's invasive nature and propensity for metastasis, the growth may be benign or malignant. Improving patient outcomes requires the early detection of malignant tumors. Despite its lower resolution and noise problems, ultrasound, which is frequently used to diagnose uterine tumors, is safer and more economical than magnetic resonance imaging (MRI) scans and biopsies. Ultrasound pictures can be processed using methods including denoising, enhancement, segmentation, and feature extraction to get around these restrictions and boost their quality. Enhanced ultrasound images can reach even higher accuracy, cementing them as plausible alternatives to MRI. A comparative analysis of copious relevant image de-speckling, image enhancement, segmentation, and feature extraction methods are carried out. Higher resolution and superior quality, strong segmented real-time ultrasound uterus tumour images are produced by using diffusion-based hybrid filters, Super Resolution Convolutional Neural Networks (SRCNN), and U-net segmentation technique. The Grey Level Co-occurrence Matrix (GLCM) and Discrete Wavelet Transform (DWT) are used to extract textural features. With the help of several machine learning approaches, such as Support Vector Machine (SVM), K-Nearest Neighbour (KNN), and Random Forest classifiers (RFC), the extracted characteristics are immediately sent to classifiers to classify uterus tumours from ultrasound images between benign and malignant. For the categorization of uterine tumors, the RFC classifier outperformed the other classifiers. The viability of cancer detection using ultrasound pictures is significantly strengthened by the suggested machine learning methodology. Multiple hospitals provided data on ultrasound pictures of uterine tumors, which were used to develop the model and obtain the prediction findings. A radiologist with 17 years of expertise in diagnostic radiology further assessed this dataset. We could produce high-quality ultrasonic real-time images of uterine tumor datasets with the suggested machine learning model at a 97.8 % accuracy rate utilizing RFC.

A Heterogeneous Group CNN for Image Super-Resolution.

Convolutional neural networks (CNNs) have obtained remarkable performance via deep architectures. However, these CNNs often achieve poor robustness for image super-resolution (SR) under complex scenes. In this article, we present a heterogeneous group SR CNN (HGSRCNN) via leveraging structure information of different types to obtain a high-quality image. Specifically, each heterogeneous group block (HGB) of HGSRCNN uses a heterogeneous architecture containing a symmetric group convolutional block and a complementary convolutional block in a parallel way to enhance the internal and external relations of different channels for facilitating richer low-frequency structure information of different types. To prevent the appearance of obtained redundant features, a refinement block (RB) with signal enhancements in a serial way is designed to filter useless information. To prevent the loss of original information, a multilevel enhancement mechanism guides a CNN to achieve a symmetric architecture for promoting expressive ability of HGSRCNN. Besides, a parallel upsampling mechanism is developed to train a blind SR model. Extensive experiments illustrate that the proposed HGSRCNN has obtained excellent SR performance in terms of both quantitative and qualitative analysis. Codes can be accessed at https://github.com/hellloxiaotian/HGSRCNN.

Learned high-resolution cardiac CT imaging from ultra-high-resolution PCD-CT.

Coronary computed tomography angiography (cCTA) is a widely used non-invasive diagnostic exam for patients with coronary artery disease (CAD). However, most clinical CT scanners are limited in spatial resolution from use of energy-integrating detectors (EIDs). Radiological evaluation of CAD is challenging, as coronary arteries are small (3-4 mm diameter) and calcifications within them are highly attenuating, leading to blooming artifacts. As such, this is a task well suited for high spatial resolution. Recently, photon-counting-detector (PCD) CT became commercially available, allowing for ultra-high resolution (UHR) data acquisition. However, PCD-CTs are costly, restricting widespread accessibility. To address this problem, we propose a super resolution convolutional neural network (CNN): ILUMENATE (Improved LUMEN visualization through Artificial super-resoluTion imagEs), creating a high resolution (HR) image simulating UHR PCD-CT. The network was trained and validated using patches extracted from 8 patients with a modified U-Net architecture. Training input and labels consisted of UHR PCD-CT images reconstructed with a smooth kernel degrading resolution (LR input) and sharp kernel (HR label). The network learned the resolution difference and was tested on 5 unseen LR patients. We evaluated network performance quantitatively and qualitatively through visual inspection, line profiles to assess spatial resolution improvements, ROIs for CT number stability and noise assessment, structural similarity index (SSIM), and percent diameter luminal stenosis. Overall, ILUMENATE improved images quantitatively and qualitatively, creating sharper edges more closely resembling reconstructed HR reference images, maintained stable CT numbers with less than 4% difference, reduced noise by 28%, maintained structural similarity (average SSIM = 0.70), and reduced percent diameter stenosis with respect to input images. ILUMENATE demonstrates potential impact for CAD patient management, improving the quality of LR CT images bringing them closer to UHR PCD-CT images.

A Combined Model of Diffusion Model and Enhanced Residual Network for Super-Resolution Reconstruction of Turbulent Flows

In this study, we introduce a novel model, the Combined Model, composed of a conditional denoising diffusion model (SR3) and an enhanced residual network (EResNet), for reconstructing high-resolution turbulent flow fields from low-resolution flow data. The SR3 model is adept at learning the distribution of flow fields. The EResNet architecture incorporates a long skip connection extending from the input directly to the output. This modification ensures the preservation of essential features learned by the SR3, while simultaneously enhancing the accuracy of the flow field. Additionally, we incorporated physical gradient constraints into the loss function of EResNet to ensure that the flow fields reconstructed by the Combined Model are consistent with the direct numerical simulation (DNS) data. Consequently, the high-resolution flow fields reconstructed by the Combined Model exhibit high conformity with the DNS results in terms of flow distribution, details, and accuracy. To validate the effectiveness of the model, experiments were conducted on two-dimensional flow around a square cylinder at a Reynolds number (Re) of 100 and turbulent channel flow at Re = 4000. The results demonstrate that the Combined Model can reconstruct both high-resolution laminar and turbulent flow fields from low-resolution data. Comparisons with a super-resolution convolutional neural network (SRCNN) and an enhanced super-resolution generative adversarial network (ESRGAN) demonstrate that while all three models perform admirably in reconstructing laminar flows, the Combined Model excels in capturing more details in turbulent flows, aligning the statistical outcomes more closely with the DNS results. Furthermore, in terms of L2 norm error, the Combined Model achieves an order of magnitude lower error compared to SRCNN and ESRGAN. Experimentation also revealed that SR3 possesses the capability to learn the distribution of flow fields. This work opens new avenues for high-fidelity flow field reconstruction using deep learning methods.

Downscaling of wind fields on the east coast of India using deep convolutional neural networks and their applications in storm surge computations

Abstract Downscaling is reconstructing data from low to high resolution, capturing local effects and magnitudes. Widely employed methods for downscaling are dynamic and statistical methods with pros and cons. With ample data, machine learning (ML) and deep learning (DL) techniques can be employed to learn mapping from low to high resolution. This article investigates convolutional neural network capabilities for downscaling winds. The speed and direction of the wind are guided by a complex relation among pressure, Coriolis force, friction, and temperature, which leads to highly nonlinear wind patterns and poses a significant challenge for downscaling. The problem can be formulated as a super-resolution technique called a super-resolution convolutional neural network (SRCNN) for data reconstruction. Few variations of SRCNN are studied for wind downscaling. Six years of European Centre for Medium-Range Weather Forecasts (ECMWF) wind datasets along the east coast of India are used in the current study and are downscaled up to four times. Downscaled winds provide better results than traditional interpolation methods. Simulations for an extreme event are conducted with SRCNN downscaled winds and are compared against interpolation methods and original data. The numerical simulation results show that DL-based methods provide results closer to the ground truth than interpolation methods.

Deep learning-based spatial downscaling and its application for tropical cyclone detection in the western North Pacific

Resolution of global climate models (GCMs) significantly influences their capacity to simulate extreme weather such as tropical cyclones (TCs). However, improving the GCM resolution is computationally expensive and time-consuming, making it challenging for many research organizations worldwide. Here, we develop a downscaling model, MSG-SE-GAN, based on the Generative Adversarial Networks (GAN) together with Multiscale Gradient (MSG) technique and a Squeeze-and-Excitation (SE) Net, to achieve 10-folded downscaling. GANs consist of a generator and a discriminator network that are trained adversarially, and are often used for generating new data that resembles a given dataset. MSG enables generation and discrimination of multi-scale images within a single model. Inclusion of an attention layer of SE captures better underlying spatial structure while preserving accuracy. The MSG-SE-GAN is stable and fast converging. It outperforms traditional bilinear interpolation and other deep-learning methods such as Super-Resolution Convolutional Neural Networks (SRCNN) and MSG-GAN in downscaling low-resolution meteorological data in assessment metrics and power spectral density. The MSG-SE-GAN has been used to downscale the TC-related variables in the western North Pacific in the low-resolution GCMs of HadGEM3-GC31 and EC-Earth3P, respectively. The downscaled data show highly similar TC activities to the direct outputs of the high-resolution HadGEM3-GC31 and EC-Earth3P, respectively. These results not only suggest the validity of the MSG-SE-GAN but also indicate its possible portability among low-resolution GCMs.

Image Super-Resolution using Convolutional Neural Networks

Image super-resolution is the process of enhancing the resolution of an image, typically from a lower resolution input to a higher resolution output. This research aims to explore the application of convolutional neural networks (CNNs) for image super-resolution. Specifically, the study will focus on developing a deep learning model capable of generating high-resolution images from low-resolution inputs. Various CNN architectures, such as SRCNN (Super-Resolution Convolutional Neural Network) or SRGAN (Super-Resolution Generative Adversarial Network), will be investigated and compared for their effectiveness in producing visually pleasing and perceptually accurate high-resolution images. Additionally, techniques such as residual learning, attention mechanisms, and adversarial training may be incorporated to further improve the quality of super-resolved images. The performance of the proposed models will be evaluated using standard image quality metrics and subjective assessments. This research has practical applications in enhancing the visual quality of low-resolution images in fields such as medical imaging, surveillance, and entertainment

Predicting the Flow Fields in Meandering Rivers with a Deep Super-Resolution Convolutional Neural Network

The hydrodynamic model, based on the strict conservation of momentum and continuity equations, can accurately simulate the distribution of a flow field. However, significant computing time and storage space requirements limit real-time prediction. Machine learning is well known for its fast computing speed and powerful learning ability, but its accuracy depends on an abundance of training data, hindering its wider use in locations without sufficient measurements. Application restrictions in data-deficient areas can be addressed through transfer learning, provided that two areas share common characteristics. In this study, a machine learning method based on a deep super-resolution convolutional neural network (DSRCNN) and transfer learning is proposed, validated, and applied to model two bend flows and one realistic test case. Firstly, the hydrodynamic model was established and validated against measured data. The validated model was considered to have the ability to generate real data and was used to generate a comprehensive data set for training and validating the machine learning model. Three different methods were compared and tested, with Realizable k-ε performing better than the others in predicting the outer bank flow distribution. DSRCNN was compared to a plain SRCNN (PSRCNN), as well as Bilinear, Nearest, and Bicubic methods, and the results showed that DSRCNN had the best performance. We compared Raw, RT, and TL methods, finding that the TL method performed the best overall. Therefore, the research results showed that the developed super-resolution convolutional neural network can provide more reliable predications and serve as an ideal tool for simulating flow field distribution in bends.

Sixty-four-fold data reduction of chest radiographs using a super-resolution convolutional neural network.

To develop and validate a super-resolution (SR) algorithm generating clinically feasible chest radiographs from 64-fold reduced data. An SR convolutional neural network was trained to produce original-resolution images (output) from 64-fold reduced images (input) using 128 × 128 patches (n = 127 030). For validation, 112 radiographs-including those with pneumothorax (n = 17), nodules (n = 20), consolidations (n = 18), and ground-glass opacity (GGO; n = 16)-were collected. Three image sets were prepared: the original images and those reconstructed using SR and conventional linear interpolation (LI) using 64-fold reduced data. The mean-squared error (MSE) was calculated to measure similarity between the reconstructed and original images, and image noise was quantified. Three thoracic radiologists evaluated the quality of each image and decided whether any abnormalities were present. The SR-images were more similar to the original images than the LI-reconstructed images (MSE: 9269 ± 1015 vs. 9429 ± 1057; P = .02). The SR-images showed lower measured noise and scored better noise level by three radiologists than both original and LI-reconstructed images (Ps < .01). The radiologists' pooled sensitivity with the SR-reconstructed images was not significantly different compared with the original images for detecting pneumothorax (SR vs. original, 90.2% [46/51] vs. 96.1% [49/51]; P = .19), nodule (90.0% [54/60] vs. 85.0% [51/60]; P = .26), consolidation (100% [54/54] vs. 96.3% [52/54]; P = .50), and GGO (91.7% [44/48] vs. 95.8% [46/48]; P = .69). SR-reconstructed chest radiographs using 64-fold reduced data showed a lower noise level than the original images, with equivalent sensitivity for detecting major abnormalities. This is the first study applying super-resolution in data reduction of chest radiographs.

Downscaling Daily Reference Evapotranspiration Using a Super-Resolution Convolutional Transposed Network

The current work proposes a novel super-resolution convolutional transposed network (SRCTN) deep learning architecture for downscaling daily climatic variables. The algorithm was established based on a super-resolution convolutional neural network with transposed convolutions. This study designed synthetic experiments to downscale daily reference evapotranspiration (ET0) data, which are a key indicator for climate change, from low resolutions (2°, 1°, and 0.5°) to a fine resolution (0.25°). The entire time period was divided into two major parts, i.e., training–validation (80%) and test periods (20%), and the training–validation period was further divided into training (80%) and validation (20%) parts. In the comparison of the downscaling performance between the SRCTN and Q-M models, the root-mean-squared error (RMSE) values indicated the accuracy of the models. For the SRCTN model, the RMSE values were reported for different scaling ratios: 0.239 for a ratio of 8, 0.077 for a ratio of 4, and 0.015 for a ratio of 2. In contrast, the RMSE values for the Q-M method were 0.334, 0.208, and 0.109 for scaling ratios of 8, 4, and 2, respectively. Notably, the RMSE values in the SRCTN model were consistently lower than those in the Q-M method across all scaling ratios, suggesting that the SRCTN model exhibited better downscaling performance in this evaluation. The results exhibited that the SRCTN method could reproduce the spatiotemporal distributions and extremes for the testing period very well. The trained SRCTN model in one study area performed remarkably well in a different area via transfer learning without re-training or calibration, and it outperformed the classic downscaling approach. The good performance of the SRCTN algorithm can be primarily attributed to the incorporation of transposed convolutions, which can be partially seen as trainable upsampling operations. Therefore, the proposed SRCTN method is a promising candidate tool for downscaling daily ET0 and can potentially be employed to conduct downscaling operations for other variables.

Channel estimation for underwater acoustic OFDM based on super‐resolution network

AbstractIn this letter, we propose a method for underwater acoustic channel estimation that combines image super‐resolution (SR) and is named FCDnNet. FCDnNet consists of two parts: Fast Super Resolution Convolutional Neural Network (FSRCNN) and Complex Denoising Convolutional Neural Network (C‐DnCNN). FSRCNN extracts effective features of pilot channels, uses deconvolution to achieve SR reconstruction, and generates a pre‐estimation channel matrix. C‐DnCNN preserves the relative positions of the real and imaginary parts of the channel, fully utilizing amplitude and phase information, and can more effectively recover the channel matrix from the pre‐estimation matrix. Experimental results show that the normalized mean square error (NMSE) of FCDnNet is at least 13.1–65.2 lower than other channel estimation methods based on deep learning.

Adaptive HEVC Video Steganography With High Performance Based on Attention-Net and PU Partition Modes

With the increasing popularity of digital video, video steganography has become a hot research topic in the field of covert communication and privacy protection. The existing prediction unit (PU) based video steganography often tends to result in large bit rate increase, which is also easily noticeable to the steganography analyst. To solve this problem, an adaptive steganography for HEVC video based on attention-net and PU partition modes is proposed. First, the distortion of modified PUs is analyzed from the perspective of rate distortion optimization at the group of pictures (GOP) level, and we find that modifying PU will lead to distortion accumulation and abnormal bitrate increase. Therefore, an adaptive distortion function based on the improved rate distortion cost is designed, and the embedding distortion is minimized by using Syndrome-Trellis Code (STC) steganography coding. Meanwhile, a super-resolution convolutional neural network with non-local sparse attention-net filter is proposed to replace the in-loop filter in HEVC to reconstruct the reference frame, thereby reducing the bitrate cost and improving the visual quality of stego-video. Experimental results show that the proposed algorithm can achieve superior perceptual quality and bitrate performance comparing with the sate-of-the-art works.

Sinogram Interpolation Inspired by Single-Image Super Resolution

Computed tomography is a medical imaging procedure used to estimate the interior of a patient or an object. Radiation scans are taken at regularly spaced angles around the object, forming a sinogram. This sinogram is then reconstructed into an image representing the contents of the object. This results in a fair amount of radiation exposure for the patient, which increases the risk of cancer. Less radiation and fewer views, however, leads to inferior image reconstruction. To solve this sparse-view problem, a deep-learning model is created that takes as input a sparse sinogram and outputs a sinogram with interpolated data for additional views. The architecture of this model is based on the super-resolution convolutional neural network. The reconstruction of modelinterpolated sinograms has less mean-squared error than the reconstruction of the sparse sinogram. It also has less mean-squared error than a reconstruction of a sinogram interpolated using the popular bilinear image-resizing algorithm. This model can be easily adapted to different image sizes, and its simplicity translates into efficiency in both time and memory requirements.

A COMPARISON OF DEEP LEARNING-BASED SUPER-RESOLUTION FRAMEWORKS FOR SENTINEL-2 IMAGERY IN URBAN AREAS

Abstract. The high-resolution images are in demand for many applications in the monitoring of urban areas. The advent of remote sensing satellites such as Sentinel-2 has made data more accessible as it provides free multispectral imagery. However, the spatial resolution of these images is not sufficient for many of the tasks. With the advent of deep learning techniques, significant progress has been made in the field of super-resolution, which has shown promising results in the improvement of the spatial resolution of satellite images. In this study, we compare four the most common deep learning-based models for the super-resolution of Sentinel-2 imagery in dense urban areas using aerial images. These methods are including enhanced deep super-resolution network (EDSR), super-resolution generative adversarial networks (ESRGAN), residual feature distillation network (RFDN), and Super-Resolution Convolutional Neural Network (SRCNN). To determine the effectiveness of the models in improving image resolution, they were evaluated using visual quality and quantitative metrics. The super-resolution results show that deep learning-based models have high potential for the generation of the high-resolution dataset from Sentinel-2 imagery in urban areas. The RFDN outperformed other deep learning-based models that achieved the peak signal-to-noise ratio (PSNR) more than 17.8.