Faster split-based feedback network for image super-resolution

Although most of the existing image super-resolution(SR)methods have achieved superior performance,contrastive learning for high-level tasks has not been fully utilized in the existing image SR methods based on deep learning.This work focuses on two well-known strategies developed for lightweight and robust SR,i.e.,contrastive learning and feedback mechanism,and proposes an integrated solution called a split-based feedback network(SPFBN).The proposed SPFBN is based on a feedback mechanism to learn abstract representations and uses contrastive learning to explore high information in the representation space.Specifically,this work first uses hidden states and constraints in recurrent neural network(RNN)to implement a feedback mechanism.Then,use contrastive learning to perform representation learning to obtain high-level information by pushing the final image to the intermediate images and pulling the final SR image to the high-resolution image.Besides,a split-based feedback block(SPFB)is proposed to reduce model redundancy,which tolerates features with similar patterns but requires fewer parameters.Extensive experimental results demonstrate the superiority of the proposed method in comparison with the state-of-the-art methods.Moreover,this work extends the experiment to prove the effectiveness of this method and shows better overall reconstruction quality.

Single frame super-resolution reconstruction based on sparse representation

In order to effectively improve the quality of recovered images, a single frame super-resolution reconstruction method based on sparse representation is proposed. The combination method of local orientation estimation-based image patch clustering and principal component analysis is used to obtain a series of geometric dictionaries of different orientations in the dictionary learning process. Subsequently, the dictionary of the nearest orientation is adaptively assigned to each of the input patches that need to be represented in the sparse coding process. Moreover, the consistency of gradients is further incorporated into the basic framework to make more substantial progress in preserving more fine edges and producing sharper results. Two groups of experiments on different types of natural images indicate that the proposed method outperforms some state-of- the-art counterparts in terms of both numerical indicators and visual quality.

Super-resolution reconstruction of synthetic-aperture radar image using adaptive-threshold singular value decomposition technique

A super-resolution reconstruction approach of （SVD） technique was presented, and its performance was radar image using an adaptive-threshold singular value decomposition analyzed, compared and assessed detailedly. First, radar imaging model and super-resolution reconstruction mechanism were outlined. Then, the adaptive-threshold SVD super-resolution algorithm, and its two key aspects, namely the determination method of point spread function （PSF） matrix T and the selection scheme of singular value threshold, were presented. Finally, the super-resolution algorithm was demonstrated successfully using the measured synthetic-aperture radar （SAR） images, and a Monte Carlo assessment was carried out to evaluate the performance of the algorithm by using the input/output signal-to-noise ratio （SNR）. Five versions of SVD algorithms, namely 1 ） using all singular values, 2） using the top 80% singular values, 3） using the top 50% singular values, 4） using the top 20% singular values and 5） using singular values s such that S2≥/max（s2）/rinsNR were tested. The experimental results indicate that when the singular value threshold is set as Smax/（rinSNR）1/2, the super-resolution algorithm provides a good compromise between too much noise and too much bias and has good reconstruction results.

Image super-resolution reconstruction based on sparse representation and residual compensation

A super-resolution reconstruction algorithm is proposed. The algorithm is based on the idea of the sparse representation of signals, by using the fact that the sparsest representation of a sig- nal is unique as the constraint of the patched-based reconstruction, and compensating residual errors of the reconstruction results both locally and globally to solve the distortion problem in patch-based reconstruction algorithms. Three reconstruction algorithms are compared. The results show that the images reconstructed with the new algorithm have the best quality.

A NOVEL ALGORITHM OF SUPER-RESOLUTION RECONSTRUCTION FOR COMPRESSED VIDEO

Super-Resolution (SR) technique means to reconstruct High-Resolution (HR) images from a sequence of Low-Resolution (LR) observations,which has been a great focus for compressed video. Based on the theory of Projection Onto Convex Set (POCS),this paper constructs Quantization Constraint Set (QCS) using the quantization information extracted from the video bit stream. By combining the statistical properties of image and the Human Visual System (HVS),a novel Adaptive Quantization Constraint Set (AQCS) is proposed. Simulation results show that AQCS-based SR al-gorithm converges at a fast rate and obtains better performance in both objective and subjective quality,which is applicable for compressed video.

Super-Resolution Image Reconstruction Based on an Improved Maximum a Posteriori Algorithm

A maximum a posteriori（ MAP） algorithm is proposed to improve the accuracy of super resolution（ SR） reconstruction in traditional methods. The algorithm applies both joints image registration and SR reconstruction in the framework,but separates them in the process of iteratiion. Firstly,we estimate the shifting parameters through two lowresolution（ LR） images and use the parameters to reconstruct initial HR images. Then,we update the shifting parameters using HR images. The aforementioned steps are repeated until the ideal HR images are obtained. The metrics such as PSNR and SSIM are used to fully evaluate the quality of the reconstructed image. Experimental results indicate that the proposed method can enhance image resolution efficiently.

Super-resolution image reconstruction based on three-step-training neural networks

A new method of super-resolution image reconstruction is proposed, which uses a three-step-training error backpropagation neural network （BPNN） to realize the super-resolution reconstruction （SRR） of satellite image. The method is based on BPNN. First, three groups learning samples with different resolutions are obtained according to image observation model, and then vector mappings are respectively used to those three group learning samples to speed up the convergence of BPNN, at last, three times consecutive training are carried on the BPNN. Training samples used in each step are of higher resolution than those used in the previous steps, so the increasing weights store a great amount of information for SRR, and network performance and generalization ability are improved greatly. Simulation and generalization tests are carried on the well-trained three-step-training NN respectively, and the reconstruction results with higher resolution images verify the effectiveness and validity of this method.

Multi-channel fast super-resolution image reconstruction based on matrix observation model

A multi-channel fast super-resolution image reconstruction algorithm based on matrix observation model is proposed in the paper,which consists of three steps to avoid the computational complexity: a single image SR reconstruction step,a registration step and a wavelet-based image fusion. This algorithm decomposes two large matrixes to the tensor product of two little matrixes and uses the natural isomorphism between matrix space and vector space to transform cost function based on matrix-vector products model to matrix form. Furthermore,we prove that the regularization part can be transformed to the matrix formed. The conjugate-gradient method is used to solve this new model. Finally,the wavelet fusion is used to integrate all the registered highresolution images obtained from the single image SR reconstruction step. The proposed algorithm reduces the storage requirement and the calculating complexity,and can be applied to large-dimension low-resolution images.

A NOVEL METHOD TO REALIZE COMPRESSED VIDEO SUPER-RESOLUTION RECONSTRUCTION

This letter proposes a novel method of compressed video super-resolution reconstruction based on MAP-POCS (Maximum Posterior Probability-Projection Onto Convex Set). At first assuming the high-resolution model subject to Poisson-Markov distribution, then constructing the projecting convex based on MAP. According to the characteristics of compressed video, two different convexes are constructed based on integrating the inter-frame and intra-frame information in the wavelet-domain. The results of the experiment demonstrate that the new method not only outperforms the traditional algorithms on the aspects of PSNR (Peak Signal-to-Noise Ratio), MSE (Mean Square Error) and reconstruction vision effect, but also has the advantages of rapid convergence and easy extension.

Super-resolution reconstruction based on CNN:A case study of Jilin-1 multispectral data

MS or MS+PAN is usually applied separately in convolutional neural network(CNN)resolution reconstruction to obtain high-resolution MS images,but the difference between the two datasets is rarely studied.This paper introduced a dual-channel network and took MS and MS+PAN of Jilin-1 spectrum satellites as two datasets to evaluate the performance of CNN resolution reconstruction,and analyzed the difference with bicubic and GS methods.The result of CNN reconstruction shows that MS+PAN dataset performed better than MS,with about 6%improvement in spatial and spectral components,and the overall quality of MS+PAN dataset was slightly higher than that of MS dataset,with QNR from 0.9559 to 0.9584.The bicubic performed best in spectral components with the quality value of 0.017,and GS performed best in spatial components with the quality values of 0.0443.CNN showed similar performance in spectral and spatial components with the two traditional methods and achieved the best overall quality with QNR value of 0.9584.

Contrastive Learning for Blind Super-Resolution via A Distortion-Specific Network

Previous deep learning-based super-resolution(SR)methods rely on the assumption that the degradation process is predefined(e.g.,bicubic downsampling).Thus,their performance would suffer from deterioration if the real degradation is not consistent with the assumption.To deal with real-world scenarios,existing blind SR methods are committed to estimating both the degradation and the super-resolved image with an extra loss or iterative scheme.However,degradation estimation that requires more computation would result in limited SR performance due to the accumulated estimation errors.In this paper,we propose a contrastive regularization built upon contrastive learning to exploit both the information of blurry images and clear images as negative and positive samples,respectively.Contrastive regularization ensures that the restored image is pulled closer to the clear image and pushed far away from the blurry image in the representation space.Furthermore,instead of estimating the degradation,we extract global statistical prior information to capture the character of the distortion.Considering the coupling between the degradation and the low-resolution image,we embed the global prior into the distortion-specific SR network to make our method adaptive to the changes of distortions.We term our distortion-specific network with contrastive regularization as CRDNet.The extensive experiments on synthetic and realworld scenes demonstrate that our lightweight CRDNet surpasses state-of-the-art blind super-resolution approaches.

Transformer and GAN-Based Super-Resolution Reconstruction Network for Medical Images

Super-resolution reconstruction in medical imaging has become more demanding due to the necessity of obtaining high-quality images with minimal radiation dose,such as in low-field magnetic resonance imaging(MRI).However,image super-resolution reconstruction remains a difficult task because of the complexity and high textual requirements for diagnosis purpose.In this paper,we offer a deep learning based strategy for reconstructing medical images from low resolutions utilizing Transformer and generative adversarial networks(T-GANs).The integrated system can extract more precise texture information and focus more on important locations through global image matching after successfully inserting Transformer into the generative adversarial network for picture reconstruction.Furthermore,we weighted the combination of content loss,adversarial loss,and adversarial feature loss as the final multi-task loss function during the training of our proposed model T-GAN.In comparison to established measures like peak signal-to-noise ratio(PSNR)and structural similarity index measure(SSIM),our suggested T-GAN achieves optimal performance and recovers more texture features in super-resolution reconstruction of MRI scanned images of the knees and belly.

松嫩沙地元素和Sr-Nd同位素组成特征及其对区域粉尘物源的指示

松嫩沙地位于欧亚黄土带最东端,其物质组成的研究有利于重建松嫩平原冰期—间冰期粉尘传输路径。为此,系统采集了松嫩沙地123个河流沙和风成沙样品,对其进行分粒级的常量、微量和稀土元素以及Sr-Nd同位素组成等地球化学分析,并利用Frequentist模型进行风尘物源定量重建,探讨松嫩沙地不同区域、不同粒级组分对哈尔滨黄土的贡献及搬运路径。结果表明,松嫩沙地经历了初级的化学风化过程(<63μm、<30μm、<10μm组分CIA平均值分别为55.20、57.46、57.51),有较低的再循环历史(<63μm、<30μm、<10μm组分CIA/WIP比值的平均值为0.98、1.08、1.04)。根据不同粒度组分的元素地球化学与Sr-Nd同位素组成特征,将松嫩沙地划分为西北部和西南部两个地球化学分区。不同粒度组分地球化学组成的定量重建结果表明,这两个分区不同粒级组分(<63μm、<30μm、<10μm)对哈尔滨黄土的贡献度分别为:75.7%~88.5%、73.4%~84.9%、61.0%~89.7%(西南部)和11.5%~24.3%、15.1%~26.6%、10.3%~39%(西北部)。本研究揭示了末次冰期以来松嫩平原粉尘传输路径以西南方向为主,与冰期以西北风为主导的环流模式存在差异。

Deep-learning-based methods for super-resolution fluorescence microscopy

The algorithm used for reconstruction or resolution enhancement is one of the factors affectingthe quality of super-resolution images obtained by fluorescence microscopy.Deep-learning-basedalgorithms have achieved stateof-the-art performance in super-resolution fluorescence micros-copy and are becoming increasingly attractive.We firstly introduce commonly-used deep learningmodels,and then review the latest applications in terms of the net work architectures,the trainingdata and the loss functions.Additionally,we discuss the challenges and limits when using deeplearning to analyze the fluorescence microscopic data,and suggest ways to improve the reliability and robustness of deep learning applications.

Research on the Application of Super Resolution Reconstruction Algorithm for Underwater Image

Underwater imaging is widely used in ocean,river and lake exploration,but it is affected by properties of water and the optics.In order to solve the lower-resolution underwater image formed by the influence of water and light,the image super-resolution reconstruction technique is applied to the underwater image processing.This paper addresses the problem of generating super-resolution underwater images by convolutional neural network framework technology.We research the degradation model of underwater images,and analyze the lower-resolution factors of underwater images in different situations,and compare different traditional super-resolution image reconstruction algorithms.We further show that the algorithm of super-resolution using deep convolution networks(SRCNN)which applied to super-resolution underwater images achieves good results.

Fast image super-resolution algorithm based on multi-resolution dictionary learning and sparse representation

Sparse representation has attracted extensive attention and performed well on image super-resolution（SR） in the last decade. However, many current image SR methods face the contradiction of detail recovery and artifact suppression. We propose a multi-resolution dictionary learning（MRDL） model to solve this contradiction, and give a fast single image SR method based on the MRDL model. To obtain the MRDL model, we first extract multi-scale patches by using our proposed adaptive patch partition method（APPM）. The APPM divides images into patches of different sizes according to their detail richness. Then, the multiresolution dictionary pairs, which contain structural primitives of various resolutions, can be trained from these multi-scale patches.Owing to the MRDL strategy, our SR algorithm not only recovers details well, with less jag and noise, but also significantly improves the computational efficiency. Experimental results validate that our algorithm performs better than other SR methods in evaluation metrics and visual perception.

A brief survey on deep learning based image super-resolution

Image super-resolution(SR)is an important technique for improving the resolution and quality of images.With the great progress of deep learning,image super-resolution achieves remarkable improvements recently.In this work,a brief survey on recent advances of deep learning based single image super-resolution methods is systematically described.The existing studies of SR techniques are roughly grouped into ten major categories.Besides,some other important issues are also introduced,such as publicly available benchmark datasets and performance evaluation metrics.Finally,this survey is concluded by highlighting four future trends.

Method of lateral image reconstruction in structured illumination microscopy with super resolution

The image reconstruction process in super-resolution structured illumination microscopy(SIM)is investigated.The structured pattern is generated by the interference of two Gaussian beams to encode undetectable spectra into detectable region of microscope.After parameters estimation of the structured pattern,the encoded spectra are computationally decoded and recombined in Fourier domain to equivalently increase the cut-off frequency of microscope,resulting in the extension of detectable spectra and a reconstructed image with about two-fold enhanced resolution.Three di®erent methods to estimate the initial phase of structured pattern are compared,verifying the auto-correlation algorithm a®ords the fast,most precise and robust measurement.The artifacts sources and detailed reconstruction°owchart for both linear and nonlinear SIM are also presented.

A lateral super-resolution imaging method using structured illumination without phase shift

Structured illumination microscopy has been a useful method for achieving lateral super-resolution,but it typically requires at least three precise phase shifts per orientation.In this paper,we propose a super-resolution method that utilizes structured illumination without phase shift.The reconstruction process requires only a conventionally illuminated image and an image with structured illumination.This method achieves the same effect as the traditional phase shift method,and more than doubles the resolution by synthesizing a few reconstructions at different illumination frequencies.We verify the resolution improvement process using a combination of theoretical derivations and diagrams,and demonstrate its effectiveness with numerical simulations.

Combination of super-resolution reconstruction and SGA-Net for marsh vegetation mapping using multi-resolution multispectral and hyperspectral images

Vegetation is crucial for wetland ecosystems.Human activities and climate changes are increasingly threatening wetland ecosystems.Combining satellite images and deep learning for classifying marsh vegetation communities has faced great challenges because of its coarse spatial resolution and limited spectral bands.This study aimed to propose a method to classify marsh vegetation using multi-resolution multispectral and hyperspectral images,combining super-resolution techniques and a novel self-constructing graph attention neural network(SGA-Net)algorithm.The SGA-Net algorithm includes a decoding layer(SCE-Net)to preciselyfine marsh vegetation classification in Honghe National Nature Reserve,Northeast China.The results indicated that the hyperspectral reconstruction images based on the super-resolution convolutional neural network(SRCNN)obtained higher accuracy with a peak signal-to-noise ratio(PSNR)of 28.87 and structural similarity(SSIM)of 0.76 in spatial quality and root mean squared error(RMSE)of 0.11 and R^(2) of 0.63 in spectral quality.The improvement of classification accuracy(MIoU)by enhanced super-resolution generative adversarial network(ESRGAN)(6.19%)was greater than that of SRCNN(4.33%)and super-resolution generative adversarial network(SRGAN)(3.64%).In most classification schemes,the SGA-Net outperformed DeepLabV3+and SegFormer algorithms for marsh vegetation and achieved the highest F1-score(78.47%).This study demonstrated that collaborative use of super-resolution reconstruction and deep learning is an effective approach for marsh vegetation mapping.