Image super-resolution reconstruction technology is currently widely used in medical imaging,video surveillance,and industrial quality inspection.It not only enhances image quality but also improves details and visual...Image super-resolution reconstruction technology is currently widely used in medical imaging,video surveillance,and industrial quality inspection.It not only enhances image quality but also improves details and visual perception,significantly increasing the utility of low-resolution images.In this study,an improved image superresolution reconstruction model based on Generative Adversarial Networks(SRGAN)was proposed.This model introduced a channel and spatial attention mechanism(CSAB)in the generator,allowing it to effectively leverage the information from the input image to enhance feature representations and capture important details.The discriminator was designed with an improved PatchGAN architecture,which more accurately captured local details and texture information of the image.With these enhanced generator and discriminator architectures and an optimized loss function design,this method demonstrated superior performance in image quality assessment metrics.Experimental results showed that this model outperforms traditional methods,presenting more detailed and realistic image details in the visual effects.展开更多
Significant advancements have been achieved in the field of Single Image Super-Resolution(SISR)through the utilization of Convolutional Neural Networks(CNNs)to attain state-of-the-art performance.Recent efforts have e...Significant advancements have been achieved in the field of Single Image Super-Resolution(SISR)through the utilization of Convolutional Neural Networks(CNNs)to attain state-of-the-art performance.Recent efforts have explored the incorporation of Transformers to augment network performance in SISR.However,the high computational cost of Transformers makes them less suitable for deployment on lightweight devices.Moreover,the majority of enhancements for CNNs rely predominantly on small spatial convolutions,thereby neglecting the potential advantages of large kernel convolution.In this paper,the authors propose a Multi-Perception Large Kernel convNet(MPLKN)which delves into the exploration of large kernel convolution.Specifically,the authors have architected a Multi-Perception Large Kernel(MPLK)module aimed at extracting multi-scale features and employ a stepwise feature fusion strategy to seamlessly integrate these features.In addition,to enhance the network's capacity for nonlinear spatial information processing,the authors have designed a Spatial-Channel Gated Feed-forward Network(SCGFN)that is capable of adapting to feature interactions across both spatial and channel dimensions.Experimental results demonstrate that MPLKN outperforms other lightweight image super-resolution models while maintaining a minimal number of parameters and FLOPs.展开更多
The application of image super-resolution(SR)has brought significant assistance in the medical field,aiding doctors to make more precise diagnoses.However,solely relying on a convolutional neural network(CNN)for image...The application of image super-resolution(SR)has brought significant assistance in the medical field,aiding doctors to make more precise diagnoses.However,solely relying on a convolutional neural network(CNN)for image SR may lead to issues such as blurry details and excessive smoothness.To address the limitations,we proposed an algorithm based on the generative adversarial network(GAN)framework.In the generator network,three different sizes of convolutions connected by a residual dense structure were used to extract detailed features,and an attention mechanism combined with dual channel and spatial information was applied to concentrate the computing power on crucial areas.In the discriminator network,using InstanceNorm to normalize tensors sped up the training process while retaining feature information.The experimental results demonstrate that our algorithm achieves higher peak signal-to-noise ratio(PSNR)and structural similarity index measure(SSIM)compared to other methods,resulting in an improved visual quality.展开更多
Gamma-ray imaging systems are powerful tools in radiographic diagnosis.However,the recorded images suffer from degradations such as noise,blurring,and downsampling,consequently failing to meet high-precision diagnosti...Gamma-ray imaging systems are powerful tools in radiographic diagnosis.However,the recorded images suffer from degradations such as noise,blurring,and downsampling,consequently failing to meet high-precision diagnostic requirements.In this paper,we propose a novel single-image super-resolution algorithm to enhance the spatial resolution of gamma-ray imaging systems.A mathematical model of the gamma-ray imaging system is established based on maximum a posteriori estimation.Within the plug-and-play framework,the half-quadratic splitting method is employed to decouple the data fidelit term and the regularization term.An image denoiser using convolutional neural networks is adopted as an implicit image prior,referred to as a deep denoiser prior,eliminating the need to explicitly design a regularization term.Furthermore,the impact of the image boundary condition on reconstruction results is considered,and a method for estimating image boundaries is introduced.The results show that the proposed algorithm can effectively addresses boundary artifacts.By increasing the pixel number of the reconstructed images,the proposed algorithm is capable of recovering more details.Notably,in both simulation and real experiments,the proposed algorithm is demonstrated to achieve subpixel resolution,surpassing the Nyquist sampling limit determined by the camera pixel size.展开更多
Due to the limitations of existing imaging hardware, obtaining high-resolution hyperspectral images is challenging. Hyperspectral image super-resolution(HSI SR) has been a very attractive research topic in computer vi...Due to the limitations of existing imaging hardware, obtaining high-resolution hyperspectral images is challenging. Hyperspectral image super-resolution(HSI SR) has been a very attractive research topic in computer vision, attracting the attention of many researchers. However, most HSI SR methods focus on the tradeoff between spatial resolution and spectral information, and cannot guarantee the efficient extraction of image information. In this paper, a multidimensional features network(MFNet) for HSI SR is proposed, which simultaneously learns and fuses the spatial,spectral, and frequency multidimensional features of HSI. Spatial features contain rich local details,spectral features contain the information and correlation between spectral bands, and frequency feature can reflect the global information of the image and can be used to obtain the global context of HSI. The fusion of the three features can better guide image super-resolution, to obtain higher-quality high-resolution hyperspectral images. In MFNet, we use the frequency feature extraction module(FFEM) to extract the frequency feature. On this basis, a multidimensional features extraction module(MFEM) is designed to learn and fuse multidimensional features. In addition, experimental results on two public datasets demonstrate that MFNet achieves state-of-the-art performance.展开更多
The employment of deep convolutional neural networks has recently contributed to significant progress in single image super-resolution(SISR)research.However,the high computational demands of most SR techniques hinder ...The employment of deep convolutional neural networks has recently contributed to significant progress in single image super-resolution(SISR)research.However,the high computational demands of most SR techniques hinder their applicability to edge devices,despite their satisfactory reconstruction performance.These methods commonly use standard convolutions,which increase the convolutional operation cost of the model.In this paper,a lightweight Partial Separation and Multiscale Fusion Network(PSMFNet)is proposed to alleviate this problem.Specifically,this paper introduces partial convolution(PConv),which reduces the redundant convolution operations throughout the model by separating some of the features of an image while retaining features useful for image reconstruction.Additionally,it is worth noting that the existing methods have not fully utilized the rich feature information,leading to information loss,which reduces the ability to learn feature representations.Inspired by self-attention,this paper develops a multiscale feature fusion block(MFFB),which can better utilize the non-local features of an image.MFFB can learn long-range dependencies from the spatial dimension and extract features from the channel dimension,thereby obtaining more comprehensive and rich feature information.As the role of the MFFB is to capture rich global features,this paper further introduces an efficient inverted residual block(EIRB)to supplement the local feature extraction ability of PSMFNet.A comprehensive analysis of the experimental results shows that PSMFNet maintains a better performance with fewer parameters than the state-of-the-art models.展开更多
Single Image Super-Resolution(SISR)technology aims to reconstruct a clear,high-resolution image with more information from an input low-resolution image that is blurry and contains less information.This technology has...Single Image Super-Resolution(SISR)technology aims to reconstruct a clear,high-resolution image with more information from an input low-resolution image that is blurry and contains less information.This technology has significant research value and is widely used in fields such as medical imaging,satellite image processing,and security surveillance.Despite significant progress in existing research,challenges remain in reconstructing clear and complex texture details,with issues such as edge blurring and artifacts still present.The visual perception effect still needs further enhancement.Therefore,this study proposes a Pyramid Separable Channel Attention Network(PSCAN)for the SISR task.Thismethod designs a convolutional backbone network composed of Pyramid Separable Channel Attention blocks to effectively extract and fuse multi-scale features.This expands the model’s receptive field,reduces resolution loss,and enhances the model’s ability to reconstruct texture details.Additionally,an innovative artifact loss function is designed to better distinguish between artifacts and real edge details,reducing artifacts in the reconstructed images.We conducted comprehensive ablation and comparative experiments on the Arabidopsis root image dataset and several public datasets.The experimental results show that the proposed PSCAN method achieves the best-known performance in both subjective visual effects and objective evaluation metrics,with improvements of 0.84 in Peak Signal-to-Noise Ratio(PSNR)and 0.017 in Structural Similarity Index(SSIM).This demonstrates that the method can effectively preserve high-frequency texture details,reduce artifacts,and have good generalization performance.展开更多
Hyperspectral images typically have high spectral resolution but low spatial resolution,which impacts the reliability and accuracy of subsequent applications,for example,remote sensingclassification and mineral identi...Hyperspectral images typically have high spectral resolution but low spatial resolution,which impacts the reliability and accuracy of subsequent applications,for example,remote sensingclassification and mineral identification.But in traditional methods via deep convolution neural net-works,indiscriminately extracting and fusing spectral and spatial features makes it challenging toutilize the differentiated information across adjacent spectral channels.Thus,we proposed a multi-branch interleaved iterative upsampling hyperspectral image super-resolution reconstruction net-work(MIIUSR)to address the above problems.We reinforce spatial feature extraction by integrat-ing detailed features from different receptive fields across adjacent channels.Furthermore,we pro-pose an interleaved iterative upsampling process during the reconstruction stage,which progres-sively fuses incremental information among adjacent frequency bands.Additionally,we add twoparallel three dimensional(3D)feature extraction branches to the backbone network to extractspectral and spatial features of varying granularity.We further enhance the backbone network’sconstruction results by leveraging the difference between two dimensional(2D)channel-groupingspatial features and 3D multi-granularity features.The results obtained by applying the proposednetwork model to the CAVE test set show that,at a scaling factor of×4,the peak signal to noiseratio,spectral angle mapping,and structural similarity are 37.310 dB,3.525 and 0.9438,respec-tively.Besides,extensive experiments conducted on the Harvard and Foster datasets demonstratethe superior potential of the proposed model in hyperspectral super-resolution reconstruction.展开更多
Due to hardware limitations,existing hyperspectral(HS)camera often suffer from low spatial/temporal resolution.Recently,it has been prevalent to super-resolve a low reso-lution(LR)HS image into a high resolution(HR)HS...Due to hardware limitations,existing hyperspectral(HS)camera often suffer from low spatial/temporal resolution.Recently,it has been prevalent to super-resolve a low reso-lution(LR)HS image into a high resolution(HR)HS image with a HR RGB(or mul-tispectral)image guidance.Previous approaches for this guided super-resolution task often model the intrinsic characteristic of the desired HR HS image using hand-crafted priors.Recently,researchers pay more attention to deep learning methods with direct supervised or unsupervised learning,which exploit deep prior only from training dataset or testing data.In this article,an efficient convolutional neural network-based method is presented to progressively super-resolve HS image with RGB image guidance.Specif-ically,a progressive HS image super-resolution network is proposed,which progressively super-resolve the LR HS image with pixel shuffled HR RGB image guidance.Then,the super-resolution network is progressively trained with supervised pre-training and un-supervised adaption,where supervised pre-training learns the general prior on training data and unsupervised adaptation generalises the general prior to specific prior for variant testing scenes.The proposed method can effectively exploit prior from training dataset and testing HS and RGB images with spectral-spatial constraint.It has a good general-isation capability,especially for blind HS image super-resolution.Comprehensive experimental results show that the proposed deep progressive learning method out-performs the existing state-of-the-art methods for HS image super-resolution in non-blind and blind cases.展开更多
Aiming at the problems such as low reconstruction efficiency,fuzzy texture details,and difficult convergence of reconstruction network face image super-resolution reconstruction algorithms,a new super-resolution recon...Aiming at the problems such as low reconstruction efficiency,fuzzy texture details,and difficult convergence of reconstruction network face image super-resolution reconstruction algorithms,a new super-resolution reconstruction algorithm with residual concern was proposed.Firstly,to solve the influence of redundant and invalid information about the face image super-resolution reconstruction network,an attention mechanism was introduced into the feature extraction module of the network,which improved the feature utilization rate of the overall network.Secondly,to alleviate the problem of gradient disappearance,the adaptive residual was introduced into the network to make the network model easier to converge during training,and features were supplemented according to the needs during training.The experimental results showed that the proposed algorithm had better reconstruction performance,more facial details,and clearer texture in the reconstructed face image than the comparison algorithm.In objective evaluation,the proposed algorithm's peak signalto-noise ratio and structural similarity were also better than other algorithms.展开更多
Super-resolution techniques are employed to enhance image resolution by reconstructing high-resolution images from one or more low-resolution inputs.Super-resolution is of paramount importance in the context of remote...Super-resolution techniques are employed to enhance image resolution by reconstructing high-resolution images from one or more low-resolution inputs.Super-resolution is of paramount importance in the context of remote sensing,satellite,aerial,security and surveillance imaging.Super-resolution remote sensing imagery is essential for surveillance and security purposes,enabling authorities to monitor remote or sensitive areas with greater clarity.This study introduces a single-image super-resolution approach for remote sensing images,utilizing deep shearlet residual learning in the shearlet transform domain,and incorporating the Enhanced Deep Super-Resolution network(EDSR).Unlike conventional approaches that estimate residuals between high and low-resolution images,the proposed approach calculates the shearlet coefficients for the desired high-resolution image using the provided low-resolution image instead of estimating a residual image between the high-and low-resolution image.The shearlet transform is chosen for its excellent sparse approximation capabilities.Initially,remote sensing images are transformed into the shearlet domain,which divides the input image into low and high frequencies.The shearlet coefficients are fed into the EDSR network.The high-resolution image is subsequently reconstructed using the inverse shearlet transform.The incorporation of the EDSR network enhances training stability,leading to improved generated images.The experimental results from the Deep Shearlet Residual Learning approach demonstrate its superior performance in remote sensing image recovery,effectively restoring both global topology and local edge detail information,thereby enhancing image quality.Compared to other networks,our proposed approach outperforms the state-of-the-art in terms of image quality,achieving an average peak signal-to-noise ratio of 35 and a structural similarity index measure of approximately 0.9.展开更多
Global and local modeling is essential for image super-resolution tasks.However,current efforts often lack explicit consideration of the cross-scale knowledge in large-scale earth observation scenarios,resulting in su...Global and local modeling is essential for image super-resolution tasks.However,current efforts often lack explicit consideration of the cross-scale knowledge in large-scale earth observation scenarios,resulting in suboptimal single-scale representations in global and local modeling.The key motivation of this work is inspired by two observations:1)There exists hierarchical features at the local and global regions in remote sensing images,and 2)they exhibit scale variation of similar ground objects(e.g.cross-scale similarity).In light of these,this paper presents an effective method to grasp the global and local image hierarchies by systematically exploring the cross-scale correlation.Specifically,we developed a Cross-scale Self-Attention(CSA)to model the global features,which introduces an auxiliary token space to calculate cross-scale self-attention matrices,thus exploring global dependency from diverse token scales.To extract the cross-scale localities,a Cross-scale Channel Attention(CCA)is devised,where multi-scale features are explored and progressively incorporated into an enriched feature.Moreover,by hierarchically deploying CSA and CCA into transformer groups,the proposed Cross-scale Hierarchical Transformer(CHT)can effectively explore cross-scale representations in remote sensing images,leading to a favorable reconstruction performance.Comprehensive experiments and analysis on four remote sensing datasets have demonstrated the superiority of CHT in both simulated and real-world remote sensing scenes.In particular,our CHT outperforms the state-of-the-art approach(TransENet)in terms of PSNR by 0.11 dB on average,but only accounts for 54.8%of its parameters.展开更多
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 l...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.展开更多
Hyperspectral image super-resolution,which refers to reconstructing the high-resolution hyperspectral image from the input low-resolution observation,aims to improve the spatial resolution of the hyperspectral image,w...Hyperspectral image super-resolution,which refers to reconstructing the high-resolution hyperspectral image from the input low-resolution observation,aims to improve the spatial resolution of the hyperspectral image,which is beneficial for subsequent applications.The development of deep learning has promoted significant progress in hyperspectral image super-resolution,and the powerful expression capabilities of deep neural networks make the predicted results more reliable.Recently,several latest deep learning technologies have made the hyperspectral image super-resolution method explode.However,a comprehensive review and analysis of the latest deep learning methods from the hyperspectral image super-resolution perspective is absent.To this end,in this survey,we first introduce the concept of hyperspectral image super-resolution and classify the methods from the perspectives with or without auxiliary information.Then,we review the learning-based methods in three categories,including single hyperspectral image super-resolution,panchromatic-based hyperspectral image super-resolution,and multispectral-based hyperspectral image super-resolution.Subsequently,we summarize the commonly used hyperspectral dataset,and the evaluations for some representative methods in three categories are performed qualitatively and quantitatively.Moreover,we briefly introduce several typical applications of hyperspectral image super-resolution,including ground object classification,urban change detection,and ecosystem monitoring.Finally,we provide the conclusion and challenges in existing learning-based methods,looking forward to potential future research directions.展开更多
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 artif...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.展开更多
Single image super resolution(SISR)is an important research content in the field of computer vision and image processing.With the rapid development of deep neural networks,different image super-resolution models have ...Single image super resolution(SISR)is an important research content in the field of computer vision and image processing.With the rapid development of deep neural networks,different image super-resolution models have emerged.Compared to some traditional SISR methods,deep learning-based methods can complete the super-resolution tasks through a single image.In addition,compared with the SISR methods using traditional convolutional neural networks,SISR based on generative adversarial networks(GAN)has achieved the most advanced visual performance.In this review,we first explore the challenges faced by SISR and introduce some common datasets and evaluation metrics.Then,we review the improved network structures and loss functions of GAN-based perceptual SISR.Subsequently,the advantages and disadvantages of different networks are analyzed by multiple comparative experiments.Finally,we summarize the paper and look forward to the future development trends of GAN-based perceptual SISR.展开更多
Although there has been a great breakthrough in the accuracy and speed of super-resolution(SR)reconstruction of a single image by using a convolutional neural network,an important problem remains unresolved:how to res...Although there has been a great breakthrough in the accuracy and speed of super-resolution(SR)reconstruction of a single image by using a convolutional neural network,an important problem remains unresolved:how to restore finer texture details during image super-resolution reconstruction?This paper proposes an Enhanced Laplacian Pyramid Generative Adversarial Network(ELSRGAN),based on the Laplacian pyramid to capture the high-frequency details of the image.By combining Laplacian pyramids and generative adversarial networks,progressive reconstruction of super-resolution images can be made,making model applications more flexible.In order to solve the problem of gradient disappearance,we introduce the Residual-in-Residual Dense Block(RRDB)as the basic network unit.Network capacity benefits more from dense connections,is able to capture more visual features with better reconstruction effects,and removes BN layers to increase calculation speed and reduce calculation complexity.In addition,a loss of content driven by perceived similarity is used instead of content loss driven by spatial similarity,thereby enhancing the visual effect of the super-resolution image,making it more consistent with human visual perception.Extensive qualitative and quantitative evaluation of the baseline datasets shows that the proposed algorithm has higher mean-sort-score(MSS)than any state-of-the-art method and has better visual perception.展开更多
Single image super-resolution(SISR)is a fundamentally challenging problem because a low-resolution(LR)image can correspond to a set of high-resolution(HR)images,while most are not expected.Recently,SISR can be achieve...Single image super-resolution(SISR)is a fundamentally challenging problem because a low-resolution(LR)image can correspond to a set of high-resolution(HR)images,while most are not expected.Recently,SISR can be achieved by a deep learning-based method.By constructing a very deep super-resolution convolutional neural network(VDSRCNN),the LR images can be improved to HR images.This study mainly achieves two objectives:image super-resolution(ISR)and deblurring the image from VDSRCNN.Firstly,by analyzing ISR,we modify different training parameters to test the performance of VDSRCNN.Secondly,we add the motion blurred images to the training set to optimize the performance of VDSRCNN.Finally,we use image quality indexes to evaluate the difference between the images from classical methods and VDSRCNN.The results indicate that the VDSRCNN performs better in generating HR images from LR images using the optimized VDSRCNN in a proper method.展开更多
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 ...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.展开更多
Convolutional neural networks(CNNs) have shown great potential for image super-resolution(SR).However,most existing CNNs only reconstruct images in the spatial domain,resulting in insufficient high-frequency details o...Convolutional neural networks(CNNs) have shown great potential for image super-resolution(SR).However,most existing CNNs only reconstruct images in the spatial domain,resulting in insufficient high-frequency details of reconstructed images.To address this issue,a channel attention based wavelet cascaded network for image super-resolution(CWSR) is proposed.Specifically,a second-order channel attention(SOCA) mechanism is incorporated into the network,and the covariance matrix normalization is utilized to explore interdependencies between channel-wise features.Then,to boost the quality of residual features,the non-local module is adopted to further improve the global information integration ability of the network.Finally,taking the image loss in the spatial and wavelet domains into account,a dual-constrained loss function is proposed to optimize the network.Experimental results illustrate that CWSR outperforms several state-of-the-art methods in terms of both visual quality and quantitative metrics.展开更多
文摘Image super-resolution reconstruction technology is currently widely used in medical imaging,video surveillance,and industrial quality inspection.It not only enhances image quality but also improves details and visual perception,significantly increasing the utility of low-resolution images.In this study,an improved image superresolution reconstruction model based on Generative Adversarial Networks(SRGAN)was proposed.This model introduced a channel and spatial attention mechanism(CSAB)in the generator,allowing it to effectively leverage the information from the input image to enhance feature representations and capture important details.The discriminator was designed with an improved PatchGAN architecture,which more accurately captured local details and texture information of the image.With these enhanced generator and discriminator architectures and an optimized loss function design,this method demonstrated superior performance in image quality assessment metrics.Experimental results showed that this model outperforms traditional methods,presenting more detailed and realistic image details in the visual effects.
文摘Significant advancements have been achieved in the field of Single Image Super-Resolution(SISR)through the utilization of Convolutional Neural Networks(CNNs)to attain state-of-the-art performance.Recent efforts have explored the incorporation of Transformers to augment network performance in SISR.However,the high computational cost of Transformers makes them less suitable for deployment on lightweight devices.Moreover,the majority of enhancements for CNNs rely predominantly on small spatial convolutions,thereby neglecting the potential advantages of large kernel convolution.In this paper,the authors propose a Multi-Perception Large Kernel convNet(MPLKN)which delves into the exploration of large kernel convolution.Specifically,the authors have architected a Multi-Perception Large Kernel(MPLK)module aimed at extracting multi-scale features and employ a stepwise feature fusion strategy to seamlessly integrate these features.In addition,to enhance the network's capacity for nonlinear spatial information processing,the authors have designed a Spatial-Channel Gated Feed-forward Network(SCGFN)that is capable of adapting to feature interactions across both spatial and channel dimensions.Experimental results demonstrate that MPLKN outperforms other lightweight image super-resolution models while maintaining a minimal number of parameters and FLOPs.
文摘The application of image super-resolution(SR)has brought significant assistance in the medical field,aiding doctors to make more precise diagnoses.However,solely relying on a convolutional neural network(CNN)for image SR may lead to issues such as blurry details and excessive smoothness.To address the limitations,we proposed an algorithm based on the generative adversarial network(GAN)framework.In the generator network,three different sizes of convolutions connected by a residual dense structure were used to extract detailed features,and an attention mechanism combined with dual channel and spatial information was applied to concentrate the computing power on crucial areas.In the discriminator network,using InstanceNorm to normalize tensors sped up the training process while retaining feature information.The experimental results demonstrate that our algorithm achieves higher peak signal-to-noise ratio(PSNR)and structural similarity index measure(SSIM)compared to other methods,resulting in an improved visual quality.
基金supported by the National Natural Science Foundation of China(Grant No.12175183)。
文摘Gamma-ray imaging systems are powerful tools in radiographic diagnosis.However,the recorded images suffer from degradations such as noise,blurring,and downsampling,consequently failing to meet high-precision diagnostic requirements.In this paper,we propose a novel single-image super-resolution algorithm to enhance the spatial resolution of gamma-ray imaging systems.A mathematical model of the gamma-ray imaging system is established based on maximum a posteriori estimation.Within the plug-and-play framework,the half-quadratic splitting method is employed to decouple the data fidelit term and the regularization term.An image denoiser using convolutional neural networks is adopted as an implicit image prior,referred to as a deep denoiser prior,eliminating the need to explicitly design a regularization term.Furthermore,the impact of the image boundary condition on reconstruction results is considered,and a method for estimating image boundaries is introduced.The results show that the proposed algorithm can effectively addresses boundary artifacts.By increasing the pixel number of the reconstructed images,the proposed algorithm is capable of recovering more details.Notably,in both simulation and real experiments,the proposed algorithm is demonstrated to achieve subpixel resolution,surpassing the Nyquist sampling limit determined by the camera pixel size.
基金supported by the Fundamental Research Funds for the Provincial Universities of Zhejiang (No.GK249909299001-036)National Key Research and Development Program of China (No. 2023YFB4502803)Zhejiang Provincial Natural Science Foundation of China (No.LDT23F01014F01)。
文摘Due to the limitations of existing imaging hardware, obtaining high-resolution hyperspectral images is challenging. Hyperspectral image super-resolution(HSI SR) has been a very attractive research topic in computer vision, attracting the attention of many researchers. However, most HSI SR methods focus on the tradeoff between spatial resolution and spectral information, and cannot guarantee the efficient extraction of image information. In this paper, a multidimensional features network(MFNet) for HSI SR is proposed, which simultaneously learns and fuses the spatial,spectral, and frequency multidimensional features of HSI. Spatial features contain rich local details,spectral features contain the information and correlation between spectral bands, and frequency feature can reflect the global information of the image and can be used to obtain the global context of HSI. The fusion of the three features can better guide image super-resolution, to obtain higher-quality high-resolution hyperspectral images. In MFNet, we use the frequency feature extraction module(FFEM) to extract the frequency feature. On this basis, a multidimensional features extraction module(MFEM) is designed to learn and fuse multidimensional features. In addition, experimental results on two public datasets demonstrate that MFNet achieves state-of-the-art performance.
基金Guangdong Science and Technology Program under Grant No.202206010052Foshan Province R&D Key Project under Grant No.2020001006827Guangdong Academy of Sciences Integrated Industry Technology Innovation Center Action Special Project under Grant No.2022GDASZH-2022010108.
文摘The employment of deep convolutional neural networks has recently contributed to significant progress in single image super-resolution(SISR)research.However,the high computational demands of most SR techniques hinder their applicability to edge devices,despite their satisfactory reconstruction performance.These methods commonly use standard convolutions,which increase the convolutional operation cost of the model.In this paper,a lightweight Partial Separation and Multiscale Fusion Network(PSMFNet)is proposed to alleviate this problem.Specifically,this paper introduces partial convolution(PConv),which reduces the redundant convolution operations throughout the model by separating some of the features of an image while retaining features useful for image reconstruction.Additionally,it is worth noting that the existing methods have not fully utilized the rich feature information,leading to information loss,which reduces the ability to learn feature representations.Inspired by self-attention,this paper develops a multiscale feature fusion block(MFFB),which can better utilize the non-local features of an image.MFFB can learn long-range dependencies from the spatial dimension and extract features from the channel dimension,thereby obtaining more comprehensive and rich feature information.As the role of the MFFB is to capture rich global features,this paper further introduces an efficient inverted residual block(EIRB)to supplement the local feature extraction ability of PSMFNet.A comprehensive analysis of the experimental results shows that PSMFNet maintains a better performance with fewer parameters than the state-of-the-art models.
基金supported by Beijing Municipal Science and Technology Project(No.Z221100007122003).
文摘Single Image Super-Resolution(SISR)technology aims to reconstruct a clear,high-resolution image with more information from an input low-resolution image that is blurry and contains less information.This technology has significant research value and is widely used in fields such as medical imaging,satellite image processing,and security surveillance.Despite significant progress in existing research,challenges remain in reconstructing clear and complex texture details,with issues such as edge blurring and artifacts still present.The visual perception effect still needs further enhancement.Therefore,this study proposes a Pyramid Separable Channel Attention Network(PSCAN)for the SISR task.Thismethod designs a convolutional backbone network composed of Pyramid Separable Channel Attention blocks to effectively extract and fuse multi-scale features.This expands the model’s receptive field,reduces resolution loss,and enhances the model’s ability to reconstruct texture details.Additionally,an innovative artifact loss function is designed to better distinguish between artifacts and real edge details,reducing artifacts in the reconstructed images.We conducted comprehensive ablation and comparative experiments on the Arabidopsis root image dataset and several public datasets.The experimental results show that the proposed PSCAN method achieves the best-known performance in both subjective visual effects and objective evaluation metrics,with improvements of 0.84 in Peak Signal-to-Noise Ratio(PSNR)and 0.017 in Structural Similarity Index(SSIM).This demonstrates that the method can effectively preserve high-frequency texture details,reduce artifacts,and have good generalization performance.
基金the National Natural Science Foun-dation of China(Nos.61471263,61872267 and U21B2024)the Natural Science Foundation of Tianjin,China(No.16JCZDJC31100)Tianjin University Innovation Foundation(No.2021XZC0024).
文摘Hyperspectral images typically have high spectral resolution but low spatial resolution,which impacts the reliability and accuracy of subsequent applications,for example,remote sensingclassification and mineral identification.But in traditional methods via deep convolution neural net-works,indiscriminately extracting and fusing spectral and spatial features makes it challenging toutilize the differentiated information across adjacent spectral channels.Thus,we proposed a multi-branch interleaved iterative upsampling hyperspectral image super-resolution reconstruction net-work(MIIUSR)to address the above problems.We reinforce spatial feature extraction by integrat-ing detailed features from different receptive fields across adjacent channels.Furthermore,we pro-pose an interleaved iterative upsampling process during the reconstruction stage,which progres-sively fuses incremental information among adjacent frequency bands.Additionally,we add twoparallel three dimensional(3D)feature extraction branches to the backbone network to extractspectral and spatial features of varying granularity.We further enhance the backbone network’sconstruction results by leveraging the difference between two dimensional(2D)channel-groupingspatial features and 3D multi-granularity features.The results obtained by applying the proposednetwork model to the CAVE test set show that,at a scaling factor of×4,the peak signal to noiseratio,spectral angle mapping,and structural similarity are 37.310 dB,3.525 and 0.9438,respec-tively.Besides,extensive experiments conducted on the Harvard and Foster datasets demonstratethe superior potential of the proposed model in hyperspectral super-resolution reconstruction.
基金National Key R&D Program of China,Grant/Award Number:2022YFC3300704National Natural Science Foundation of China,Grant/Award Numbers:62171038,62088101,62006023。
文摘Due to hardware limitations,existing hyperspectral(HS)camera often suffer from low spatial/temporal resolution.Recently,it has been prevalent to super-resolve a low reso-lution(LR)HS image into a high resolution(HR)HS image with a HR RGB(or mul-tispectral)image guidance.Previous approaches for this guided super-resolution task often model the intrinsic characteristic of the desired HR HS image using hand-crafted priors.Recently,researchers pay more attention to deep learning methods with direct supervised or unsupervised learning,which exploit deep prior only from training dataset or testing data.In this article,an efficient convolutional neural network-based method is presented to progressively super-resolve HS image with RGB image guidance.Specif-ically,a progressive HS image super-resolution network is proposed,which progressively super-resolve the LR HS image with pixel shuffled HR RGB image guidance.Then,the super-resolution network is progressively trained with supervised pre-training and un-supervised adaption,where supervised pre-training learns the general prior on training data and unsupervised adaptation generalises the general prior to specific prior for variant testing scenes.The proposed method can effectively exploit prior from training dataset and testing HS and RGB images with spectral-spatial constraint.It has a good general-isation capability,especially for blind HS image super-resolution.Comprehensive experimental results show that the proposed deep progressive learning method out-performs the existing state-of-the-art methods for HS image super-resolution in non-blind and blind cases.
基金supported by National Natural Science Foundation of China(No.62063014)。
文摘Aiming at the problems such as low reconstruction efficiency,fuzzy texture details,and difficult convergence of reconstruction network face image super-resolution reconstruction algorithms,a new super-resolution reconstruction algorithm with residual concern was proposed.Firstly,to solve the influence of redundant and invalid information about the face image super-resolution reconstruction network,an attention mechanism was introduced into the feature extraction module of the network,which improved the feature utilization rate of the overall network.Secondly,to alleviate the problem of gradient disappearance,the adaptive residual was introduced into the network to make the network model easier to converge during training,and features were supplemented according to the needs during training.The experimental results showed that the proposed algorithm had better reconstruction performance,more facial details,and clearer texture in the reconstructed face image than the comparison algorithm.In objective evaluation,the proposed algorithm's peak signalto-noise ratio and structural similarity were also better than other algorithms.
文摘Super-resolution techniques are employed to enhance image resolution by reconstructing high-resolution images from one or more low-resolution inputs.Super-resolution is of paramount importance in the context of remote sensing,satellite,aerial,security and surveillance imaging.Super-resolution remote sensing imagery is essential for surveillance and security purposes,enabling authorities to monitor remote or sensitive areas with greater clarity.This study introduces a single-image super-resolution approach for remote sensing images,utilizing deep shearlet residual learning in the shearlet transform domain,and incorporating the Enhanced Deep Super-Resolution network(EDSR).Unlike conventional approaches that estimate residuals between high and low-resolution images,the proposed approach calculates the shearlet coefficients for the desired high-resolution image using the provided low-resolution image instead of estimating a residual image between the high-and low-resolution image.The shearlet transform is chosen for its excellent sparse approximation capabilities.Initially,remote sensing images are transformed into the shearlet domain,which divides the input image into low and high frequencies.The shearlet coefficients are fed into the EDSR network.The high-resolution image is subsequently reconstructed using the inverse shearlet transform.The incorporation of the EDSR network enhances training stability,leading to improved generated images.The experimental results from the Deep Shearlet Residual Learning approach demonstrate its superior performance in remote sensing image recovery,effectively restoring both global topology and local edge detail information,thereby enhancing image quality.Compared to other networks,our proposed approach outperforms the state-of-the-art in terms of image quality,achieving an average peak signal-to-noise ratio of 35 and a structural similarity index measure of approximately 0.9.
基金supported in part by the National Natural Science Foundation of China[grant numbers 42230108,and 61971319].
文摘Global and local modeling is essential for image super-resolution tasks.However,current efforts often lack explicit consideration of the cross-scale knowledge in large-scale earth observation scenarios,resulting in suboptimal single-scale representations in global and local modeling.The key motivation of this work is inspired by two observations:1)There exists hierarchical features at the local and global regions in remote sensing images,and 2)they exhibit scale variation of similar ground objects(e.g.cross-scale similarity).In light of these,this paper presents an effective method to grasp the global and local image hierarchies by systematically exploring the cross-scale correlation.Specifically,we developed a Cross-scale Self-Attention(CSA)to model the global features,which introduces an auxiliary token space to calculate cross-scale self-attention matrices,thus exploring global dependency from diverse token scales.To extract the cross-scale localities,a Cross-scale Channel Attention(CCA)is devised,where multi-scale features are explored and progressively incorporated into an enriched feature.Moreover,by hierarchically deploying CSA and CCA into transformer groups,the proposed Cross-scale Hierarchical Transformer(CHT)can effectively explore cross-scale representations in remote sensing images,leading to a favorable reconstruction performance.Comprehensive experiments and analysis on four remote sensing datasets have demonstrated the superiority of CHT in both simulated and real-world remote sensing scenes.In particular,our CHT outperforms the state-of-the-art approach(TransENet)in terms of PSNR by 0.11 dB on average,but only accounts for 54.8%of its parameters.
基金the National Key R&D Program of China(No.2019YFB1405900)the National Natural Science Foundation of China(No.62172035,61976098)。
文摘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.
基金supported in part by the National Natural Science Foundation of China(62276192)。
文摘Hyperspectral image super-resolution,which refers to reconstructing the high-resolution hyperspectral image from the input low-resolution observation,aims to improve the spatial resolution of the hyperspectral image,which is beneficial for subsequent applications.The development of deep learning has promoted significant progress in hyperspectral image super-resolution,and the powerful expression capabilities of deep neural networks make the predicted results more reliable.Recently,several latest deep learning technologies have made the hyperspectral image super-resolution method explode.However,a comprehensive review and analysis of the latest deep learning methods from the hyperspectral image super-resolution perspective is absent.To this end,in this survey,we first introduce the concept of hyperspectral image super-resolution and classify the methods from the perspectives with or without auxiliary information.Then,we review the learning-based methods in three categories,including single hyperspectral image super-resolution,panchromatic-based hyperspectral image super-resolution,and multispectral-based hyperspectral image super-resolution.Subsequently,we summarize the commonly used hyperspectral dataset,and the evaluations for some representative methods in three categories are performed qualitatively and quantitatively.Moreover,we briefly introduce several typical applications of hyperspectral image super-resolution,including ground object classification,urban change detection,and ecosystem monitoring.Finally,we provide the conclusion and challenges in existing learning-based methods,looking forward to potential future research directions.
文摘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.
基金The authors are highly thankful to the Development Research Center of Guangxi Relatively Sparse-populated Minorities(ID:GXRKJSZ201901)to the Natural Science Foundation of Guangxi Province(No.2018GXNSFAA281164)This research was financially supported by the project of outstanding thousand young teachers’training in higher education institutions of Guangxi,Guangxi Colleges and Universities Key Laboratory Breeding Base of System Control and Information Processing.
文摘Single image super resolution(SISR)is an important research content in the field of computer vision and image processing.With the rapid development of deep neural networks,different image super-resolution models have emerged.Compared to some traditional SISR methods,deep learning-based methods can complete the super-resolution tasks through a single image.In addition,compared with the SISR methods using traditional convolutional neural networks,SISR based on generative adversarial networks(GAN)has achieved the most advanced visual performance.In this review,we first explore the challenges faced by SISR and introduce some common datasets and evaluation metrics.Then,we review the improved network structures and loss functions of GAN-based perceptual SISR.Subsequently,the advantages and disadvantages of different networks are analyzed by multiple comparative experiments.Finally,we summarize the paper and look forward to the future development trends of GAN-based perceptual SISR.
基金This work was supported in part by the National Science Foundation of China under Grant 61572526.
文摘Although there has been a great breakthrough in the accuracy and speed of super-resolution(SR)reconstruction of a single image by using a convolutional neural network,an important problem remains unresolved:how to restore finer texture details during image super-resolution reconstruction?This paper proposes an Enhanced Laplacian Pyramid Generative Adversarial Network(ELSRGAN),based on the Laplacian pyramid to capture the high-frequency details of the image.By combining Laplacian pyramids and generative adversarial networks,progressive reconstruction of super-resolution images can be made,making model applications more flexible.In order to solve the problem of gradient disappearance,we introduce the Residual-in-Residual Dense Block(RRDB)as the basic network unit.Network capacity benefits more from dense connections,is able to capture more visual features with better reconstruction effects,and removes BN layers to increase calculation speed and reduce calculation complexity.In addition,a loss of content driven by perceived similarity is used instead of content loss driven by spatial similarity,thereby enhancing the visual effect of the super-resolution image,making it more consistent with human visual perception.Extensive qualitative and quantitative evaluation of the baseline datasets shows that the proposed algorithm has higher mean-sort-score(MSS)than any state-of-the-art method and has better visual perception.
文摘Single image super-resolution(SISR)is a fundamentally challenging problem because a low-resolution(LR)image can correspond to a set of high-resolution(HR)images,while most are not expected.Recently,SISR can be achieved by a deep learning-based method.By constructing a very deep super-resolution convolutional neural network(VDSRCNN),the LR images can be improved to HR images.This study mainly achieves two objectives:image super-resolution(ISR)and deblurring the image from VDSRCNN.Firstly,by analyzing ISR,we modify different training parameters to test the performance of VDSRCNN.Secondly,we add the motion blurred images to the training set to optimize the performance of VDSRCNN.Finally,we use image quality indexes to evaluate the difference between the images from classical methods and VDSRCNN.The results indicate that the VDSRCNN performs better in generating HR images from LR images using the optimized VDSRCNN in a proper method.
基金the National Key Research and Development Program of China(No.2019YFB1405900)。
文摘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.
基金Supported by the National Natural Science Foundation of China(No.61901183)Fundamental Research Funds for the Central Universities(No.ZQN921)+4 种基金Natural Science Foundation of Fujian Province Science and Technology Department(No.2021H6037)Key Project of Quanzhou Science and Technology Plan(No.2021C008R)Natural Science Foundation of Fujian Province(No.2019J01010561)Education and Scientific Research Project for Young and Middle-aged Teachers of Fujian Province 2019(No.JAT191080)Science and Technology Bureau of Quanzhou(No.2017G046)。
文摘Convolutional neural networks(CNNs) have shown great potential for image super-resolution(SR).However,most existing CNNs only reconstruct images in the spatial domain,resulting in insufficient high-frequency details of reconstructed images.To address this issue,a channel attention based wavelet cascaded network for image super-resolution(CWSR) is proposed.Specifically,a second-order channel attention(SOCA) mechanism is incorporated into the network,and the covariance matrix normalization is utilized to explore interdependencies between channel-wise features.Then,to boost the quality of residual features,the non-local module is adopted to further improve the global information integration ability of the network.Finally,taking the image loss in the spatial and wavelet domains into account,a dual-constrained loss function is proposed to optimize the network.Experimental results illustrate that CWSR outperforms several state-of-the-art methods in terms of both visual quality and quantitative metrics.
