Noise robust single image super-resolution using a multiscale image pyramid

In this paper, we proposed improved single image super resolution using multiple re-evaluation Process model for use in various image processing fields. The proposed method generates the first super resolution using the input image, and analyzes the change for each region by comparing the features of previous image and super resolution result. According to the analyzed features, the feature map for generate n-th super resolution was selected for improved detail. After then, next generate super resolution using previous super resolution result as input image. This process is repeated for final result. The existing single image super resolution method has two areas to be improved. First, it minimizes artifacts or staircases, which are unnecessary details that can be created during the super resolution process. Second, it is necessary to consider the input image because it affects the result depending on the quality of input image used in the super resolution process. Therefore, in order to minimize unnecessary details, the proposed method analyzed the feature map from the generated super resolution result and applied it according to the amount of change. In addition, aimed to gradually improve the input image to be used in the super resolution process by using the super resolution generated in the previous step. By comparing and evaluating the proposed method with the conventional single image super resolution method with PSNR and SSIM, it is improved by about 3%.

Image super resolution is a vital research topic in the field of computer vision. It aims to reconstruct high resolution images from low resolution images. Although the conventional image super resolution methods have achieved good performance and effect, there are still have some issues, e.g., the high-frequency details information is insufficient, and the reconstruction process will bring additional noise, and most basic interpolation techniques produce blurry results. To settle the problems mentioned above, we consider combining the deep learning method with the frequency domain fusion method. In this paper, a novel single image super resolution method based on SRFeat network and wavelet fusion is proposed. First, the training image is taken as the input of the backbone SRFeat network, then the generative adversarial network training is carried out. Then, the up-sampling is utilized to obtain the coarse super resolved image. Finally, the output image after the network training is combined with the up-sampling image of the low-resolution image by Wavelet fusion to obtain the final result. Without increasing the depth of the network and the redundant parameters, the proposed method can achieve better reconstruct result. The experimental results show that the proposed method can not only reduce the probability of image distortion, but recover the global information of the reconstructed image and remove the noise brought by the reconstruction process. The PSNR value of the proposed method is improved 0.3 dB, and the SSIM is improved 0.02.

Single image super-resolution (SISR) has developed fast for recent years. Most of the SISR models are trained and evaluated with simulated data where low-resolution (LR) images are generated from high-resolution (HR) images using pre-defined degradation. In contrast, real-world image super-resolution (RealSR) is more challenging since the process of obtaining LR images is formulated by complex degradation. To solve this problem, we propose the multi-scale adaptive real image super-resolution (MARS). Our model extracts complex features in the image and uses them for upscaling adaptively. Experimental results show that the proposed method can improve the quality of the super-resolved images in RealSR.

We present a self-learning approach for single image super-resolution (SR), with the ability to preserve high frequency components such as edges in resulting high resolution (HR) images. Given a low-resolution (LR) input image, we construct its image pyramid and produce a super pixel dataset. By extracting context information from the super-pixels, we propose to deploy context-specific contour let transform on them in order to model the relationship (via support vector regression) between the input patches and their associated directional high-frequency responses. These learned models are applied to predict the SR output with satisfactory quality. Unlike prior learning-based SR methods, our approach advances a self-learning technique and does not require the self similarity of image patches within or across image scales. More importantly, we do not need to collect training LR/HR image data in advance and only require a single LR input image. Empirical results verify the effectiveness of our approach, which quantitatively and qualitatively outperforms existing interpolation or learning-based SR methods.

Single image super-resolution (SISR) is of great importance as a low-level computer vision task. The fast development of Generative Adversarial Network (GAN) based deep learning architectures realises an efficient and effective SISR to boost the spatial resolution of natural images captured by digital cameras. However, the SISR for medical images is still a very challenging problem. This is due to (1) compared to natural images, in general, medical images have lower signal to noise ratios, (2) GAN based models pre-trained on natural images may synthesise unrealistic patterns in medical images which could affect the clinical interpretation and diagnosis, and (3) the vanilla GAN architecture may suffer from unstable training and collapse mode that can also affect the SISR results. In this paper, we propose a novel lesion focused SR (LFSR) method, which incorporates GAN to achieve perceptually realistic SISR results for brain tumour MRI images. More importantly, we test and make comparison using recently developed GAN variations, e.g., Wasserstein GAN (WGAN) and WGAN with Gradient Penalty (WGAN-GP), and propose a novel multi-scale GAN (MS-GAN), to achieve a more stabilised and efficient training and improved perceptual quality of the super-resolved results. Based on both quantitative evaluations and our designed mean opinion score, the proposed LFSR coupled with MS-GAN has performed better in terms of both perceptual quality and efficiency.

Recently self-similarity has been used for super resolution which generates favorable results. In this paper, single image super resolution method using self-example-based method is proposed. Patch redundancy cross-scale images is fully considered and patch similarity in image pyramids is used to improve the image resolution. Also the local structural constraints with steering kernel regression for patch similarity are used in the image reconstruction. For avoiding over-smoothing the structure of image, an automatic metric is presented to preserve the structure better. The patch self-similarity and local structure regularity in the image pyramids are combined to get the high resolution image. The results show that the proposed method has higher quality as compared to other state-of-art super resolution methods.

MSAR-Net: Multi-scale attention based light-weight image super-resolution

Special Section on Enabling Technologies for Network-Based Applications

The single image super resolution recovers missing high resolution details so as to reconstruct a high resolution image from a single low resolution image. This paper proposes a novel directionally adaptive, learning-based, single image super resolution method using multiple direction wavelet transform, called directionlets. Here, critically sampled directionlets are used to capture directional features effectively and to extract edge information along different directions from a set of available high resolution images. This information is used as the training set for super resolving a low resolution input image. The directionlet coefficients at finer scales of its high resolution image are learned locally from this training set and the inverse directionlet transform recovers the super resolved high resolution image. The simulation results showed that the proposed directionlet approach outperforms standard interpolation techniques like cubic spline interpolation as well as standard wavelet-based learning, both visually and in terms of the mean squared error (MSE) values. The SNR scores for cubic spline interpolation, wavelet and directionlet method are 13.6998 dB, 23.8324 dB and 30.8654 dB respectively for Barbara.

Lightweight single image super-resolution with attentive residual refinement network

High resolution images obtained from single image super resolution often suffer from excessive smoothness which blurs edges and gives smooth areas an un-natural texture. This paper proposes a super resolution method that combines wavelet super resolution and an edge enhancing back-projection filter to produce a super resolution image with sharp edges and natural texture. The low resolution image is first up-sampled using the discrete wavelet transform. High frequency sub-bands are derived from DWT of a bicubic interpolated low resolution image. The up-sampled image is then subjected to iterative back projection using a back-projection filter based on the Laplacian of Gaussian. The proposed method is tested on RGB colour images. It is found to produce better results than bicubic interpolation and selected edge directed interpolation methods in terms of PSNR and SSIM. It is also found to approximate the edge sharpness of edge directed interpolation while giving a more natural texture in smooth areas. The proposed method also eliminates spurious colours that have been observed in most edge-directed methods.

To study the effect of decentration and tilt of the type I Boston keratoprosthesis (KPro) on image quality in both aphakic and pseudophakic eyes. An optical ray-tracing program was used to simulate the image projected onto the retina in an eye with a perfectly centered KPro, and in eyes with varying degrees of KPro decentration and tilt. Decentration was modeled along a typical white-to-white distance of 12.0 mm, and the corresponding tilt was calculated assuming a radius of curvature of 8.0 mm, the radius of curvature of the backplate of the KPro. Both aphakic and pseudophakic eyes were simulated, and the corresponding modulation transfer function curves, point spread functions, and astigmatism were analyzed. The perfectly centered KPro produced a high-quality image with no induced astigmatism. Increasing decentration beyond approximately 0.5 mm resulted in poorer image quality with a more pronounced effect in the presence of an intraocular lens. Using models of the normal eye as a threshold, image degradation due to decentration becomes clinically significant at approximately 1.4 mm and 0.9 mm for the aphakic and pseudophakic cases, respectively. Astigmatism introduced by decentration is approximately 0.25 D cylinders at those thresholds. Decentration of up to 0.5 mm had no significant impact on image quality and an attempt at good intraoperative centration of the KPro within this range is important. Conversely, decentration of 0.9 mm or more during surgical implantation can result in significant degradation in retinal image quality including astigmatism. The effect is greater in the pseudophakic eye.

ObjectivesTo evaluate the radiation exposure, quantitative, and qualitative image quality in pediatric cardiac CT by using photon-counting detector computed tomography (PCD CT) versus energy-integrating detector CT (EID CT) in matched children.Materials and methodsThirty-seven contrast-enhanced, clinically indicated cardiac CTs performed on PCD CT were matched with 37 examinations acquired by EID CT. The patients were matched according to water-equivalent diameters. Quantitative evaluation of image quality comprised a region of interest (ROI)-based analysis, calculating image noise, signal-to-noise (SNR) and contrast-to-noise (CNR) ratio. Differences of the attenuation variation of the paraspinal and the pectoral muscles were calculated to measure beam hardening artifacts. Volume CT dose index (CTDIvol) and dose length product (DLP) were documented, and the effective radiation dose was calculated for each patient. Statistical analysis comprised t-tests and Wilcoxon signed rank tests.ResultsThe mean age of the children on PCD CT was 794 ± 1016 days, similar to the mean age of 815 ± 957 days of the children on EID CT (p = 0.76). Moreover, age, height, weight, and body mass index (BMI) were also not significantly different between the two groups (p ≥ 0.32). Radiation exposure was significantly lower on PCD CT (CTDIvol 0.20 ± 0.12 mGy and DLP 4.06 ± 3.22 mGy*cm) versus EID CT (CTDIvol 0.37 ± 0.17 mGy, p < 0.001 and DLP 7.21 ± 4.67 mGy*cm, p < 0.001). No significant differences in SNR, CNR, or beam hardening artifacts could be observed. Qualitative image quality was also comparable for PCD CT versus EID CT.ConclusionsWith a reduction in radiation exposure exceeding 40% by using PCD CT, image quality remained stable compared to EID CT. Reducing radiation with PCD CT while preserving image quality might substantially advance cardiac imaging in children.Key PointsQuestionChildren are particularly sensitive to radiation exposure, highlighting the need for dose reduction.FindingsRadiation dosage can be significantly reduced while preserving image quality when using photon-counting detector (PCD) CT in pediatric patients with congenital heart disease.Clinical relevanceSince radiation exposure can be significantly reduced by PCD CT compared to energy-integrating detector (EID) CT, while image quality was comparable, PCD CT is advisable for children with congenital heart disease.Graphical

The resolution of remote-sensing images is directly related to the application value of data products. Due to differences in imaging characteristics between digital cameras and remote-sensing cameras, the existing network models cannot get the best super resolution (SR) results of satellite images. In response to the requirements of remote-sensing image production, we propose a single image super-resolution (SISR) reconstruction method for specific type of remote-sensing satellite. First, we measure and model the imaging degradation phenomenon of remote-sensing satellite, including the image blur and noise model. Then, high-quality aerial images are down-sampled and degraded to construct paired training image datasets. We chose the most popular Enhanced Super-Resolution Generative Adversarial Networks (ESRGAN) as the basic structure and optimized the number of Residual-in-Residual Dense Block (RRDB) modules to further improve the processing efficiency. Finally, we perform a series of quantitative measurements of the SR image results, including image interpretation capability, reconstruction accuracy, ground resolution distance, and data processing efficiency, using higher resolution remote-sensing images as the benchmark. The experimental results demonstrate that our proposed method has higher interpretation capability and reconstruction accuracy for the SR processing of specific type remote-sensing satellite. Our proposed method is evaluated within a real satellite image product, that demonstrated it has the capability of pipeline super-resolution processing of remote sensing data products.

In recent years, the study of a single image super-resolution (SISR) is crucial to improving image resolution and using hardware technology to improve image resolution. SISR is widely used in satellite remote sensing, video surveillance, and medical image processing because it mainly relies on deep learning algorithms to realize the conversion from low-resolution (LR) images to high-resolution images. It has the advantages of low cost, simple operation, and high efficiency. This paper proposes an image super-resolution method based on a generative adversarial network named text localization generative adversarial nets (TLGAN) model. The method is improved based on super-resolution generative adversarial networks (SRGAN), and the batch normalization layer is removed, which significantly reduces the computational burden of the model. In TLGAN model, we used the transfer learning method to pre-trained the model on the large dataset ImageNet, and then apply the pre-trained model to the cartoon image data set animes to achieve image super-resolution. Experimental results report that the proposed method has the advantages of fast running speed and excellent visual perception of super-resolution images compared with bicubic interpolation and SRGAN method.