RESUMEN

Vision Transformer have achieved impressive performance in image super-resolution. However, they suffer from low inference speed mainly because of the quadratic complexity of multi-head self-attention (MHSA), which is the key to learning long-range dependencies. On the contrary, most CNN-based methods neglect the important effect of global contextual information, resulting in inaccurate and blurring details. If one can make the best of both Transformers and CNNs, it will achieve a better trade-off between image quality and inference speed. Based on this observation, firstly assume that the main factor affecting the performance in the Transformer-based SR models is the general architecture design, not the specific MHSA component. To verify this, some ablation studies are made by replacing MHSA with large kernel convolutions, alongside other essential module replacements. Surprisingly, the derived models achieve competitive performance. Therefore, a general architecture design GlobalSR is extracted by not specifying the core modules including blocks and domain embeddings of Transformer-based SR models. It also contains three practical guidelines for designing a lightweight SR network utilizing image-level global contextual information to reconstruct SR images. Following the guidelines, the blocks and domain embeddings of GlobalSR are instantiated via Deformable Convolution Attention Block (DCAB) and Fast Fourier Convolution Domain Embedding (FCDE), respectively. The instantiation of general architecture, termed GlobalSR-DF, proposes a DCA to extract the global contextual feature by utilizing Deformable Convolution and a Hadamard product as the attention map at the block level. Meanwhile, the FCDE utilizes the Fast Fourier to transform the input spatial feature into frequency space and then extract image-level global information from it by convolutions. Extensive experiments demonstrate that GlobalSR is the key part in achieving a superior trade-off between SR quality and efficiency. Specifically, our proposed GlobalSR-DF outperforms state-of-the-art CNN-based and ViT-based SISR models regarding accuracy-speed trade-offs with sharp and natural details.

RESUMEN

Gleason grading system is dependable for quantifying prostate cancer. This paper introduces a fast multiphoton microscopic imaging method via deep learning for automatic Gleason grading. Due to the contradiction between multiphoton microscopy (MPM) imaging speed and quality, a deep learning architecture (SwinIR) is used for image super-resolution to address this issue. The quality of low-resolution image is improved, which increased the acquisition speed from 7.55 s per frame to 0.24 s per frame. A classification network (Swin Transformer) was introduced for automated Gleason grading. The classification accuracy and Macro-F1 achieved by training on high-resolution images are respectively 90.9% and 90.9%. For training on super-resolution images, the classification accuracy and Macro-F1 are respectively 89.9% and 89.9%. It shows that super-resolution image can provide a comparable performance to high-resolution image. Our results suggested that MPM joint image super-resolution and automatic classification methods hold the potential to be a real-time clinical diagnostic tool for prostate cancer diagnosis.

RESUMEN

The success of the ClassSR has led to a strategy of decomposing images being used for large image SR. The decomposed image patches have different recovery difficulties. Therefore, in ClassSR, image patches are reconstructed by different networks to greatly reduce the computational cost. However, in ClassSR, the training of multiple sub-networks inevitably increases the training difficulty. Furthermore, decomposing images with overlapping not only increases the computational cost but also inevitably produces artifacts. To address these challenges, we propose an end-to-end general framework, named patches separation and artifacts removal SR (PSAR-SR). In PSAR-SR, we propose an image information complexity module (IICM) to efficiently determine the difficulty of recovering image patches. Then, we propose a patches classification and separation module (PCSM), which can dynamically select an appropriate SR path for image patches of different recovery difficulties. Moreover, we propose a multi-attention artifacts removal module (MARM) in the network backend, which can not only greatly reduce the computational cost but also solve the artifacts problem well under the overlapping-free decomposition. Further, we propose two loss functions - threshold penalty loss (TP-Loss) and artifacts removal loss (AR-Loss). TP-Loss can better select appropriate SR paths for image patches. AR-Loss can effectively guarantee the reconstruction quality between image patches. Experiments show that compared to the leading methods, PSAR-SR well eliminates artifacts under the overlapping-free decomposition and achieves superior performance on existing methods (e.g., FSRCNN, CARN, SRResNet, RCAN and CAMixerSR). Moreover, PSAR-SR saves 53%-65% FLOPs in computational cost far beyond the leading methods. The code will be made available: https://github.com/dywang95/PSAR-SR.

Asunto(s)

RESUMEN

In the field of endoscopic imaging, challenges such as low resolution, complex textures, and blurred edges often degrade the quality of 3D reconstructed models. To address these issues, this study introduces an innovative endoscopic image super-resolution and 3D reconstruction technique named Omni-Directional Focus and Scale Resolution (OmDF-SR). This method integrates an Omnidirectional Self-Attention (OSA) mechanism, an Omnidirectional Scale Aggregation Group (OSAG), a Dual-stream Adaptive Focus Mechanism (DAFM), and a Dynamic Edge Adjustment Framework (DEAF) to enhance the accuracy and efficiency of super-resolution processing. Additionally, it employs Structure from Motion (SfM) and Multi-View Stereo (MVS) technologies to achieve high-precision medical 3D models. Experimental results indicate significant improvements in image processing with a PSNR of 38.2902 dB and an SSIM of 0.9746 at a magnification factor of ×2, and a PSNR of 32.1723 dB and an SSIM of 0.9489 at ×4. Furthermore, the method excels in reconstructing detailed 3D models, enhancing point cloud density, mesh quality, and texture mapping richness, thus providing substantial support for clinical diagnosis and surgical planning.

RESUMEN

In this paper, we propose a Local Global Union Network (LGUN), which effectively combines the strengths of Transformers and Convolutional Networks to develop a lightweight and high-performance network suitable for Single Image Super-Resolution (SISR). Specifically, we make use of the advantages of Transformers to provide input-adaptation weighting and global context interaction. We also make use of the advantages of Convolutional Networks to include spatial inductive biases and local connectivity. In the shallow layer, the local spatial information is encoded by Multi-order Local Hierarchical Attention (MLHA). In the deeper layer, we utilize Dynamic Global Sparse Attention (DGSA), which is based on the Multi-stage Token Selection (MTS) strategy to model global context dependencies. Moreover, we also conduct extensive experiments on both natural and satellite datasets, acquired through optical and satellite sensors, respectively, demonstrating that LGUN outperforms existing methods.

RESUMEN

Infrared images hold significant value in applications such as remote sensing and fire safety. However, infrared detectors often face the problem of high hardware costs, which limits their widespread use. Advancements in deep learning have spurred innovative approaches to image super-resolution (SR), but comparatively few efforts have been dedicated to the exploration of infrared images. To address this, we design the Residual Swin Transformer and Average Pooling Block (RSTAB) and propose the SwinAIR, which can effectively extract and fuse the diverse frequency features in infrared images and achieve superior SR reconstruction performance. By further integrating SwinAIR with U-Net, we propose the SwinAIR-GAN for real infrared image SR reconstruction. SwinAIR-GAN extends the degradation space to better simulate the degradation process of real infrared images. Additionally, it incorporates spectral normalization, dropout, and artifact discrimination loss to reduce the potential image artifacts. Qualitative and quantitative evaluations on various datasets confirm the effectiveness of our proposed method in reconstructing realistic textures and details of infrared images.

RESUMEN

3D digital-image correlation (3D-DIC) is a non-contact optical technique for full-field shape, displacement, and deformation measurement. Given the high experimental hardware costs associated with 3D-DIC, the development of high-fidelity 3D-DIC simulations holds significant value. However, existing research on 3D-DIC simulation was mainly carried out through the generation of random speckle images. This study innovatively proposes a complete 3D-DIC simulation method involving optical simulation and mechanical simulation and integrating 3D-DIC, virtual stereo vision, and image super-resolution reconstruction technology. Virtual stereo vision can reduce hardware costs and eliminate camera-synchronization errors. Image super-resolution reconstruction can compensate for the decrease in precision caused by image-resolution loss. An array of software tools such as ANSYS SPEOS 2024R1, ZEMAX 2024R1, MECHANICAL 2024R1, and MULTIDIC v1.1.0 are used to implement this simulation. Measurement systems based on stereo vision and virtual stereo vision were built and tested for use in 3D-DIC. The results of the simulation experiment show that when the synchronization error of the basic stereo-vision system (BSS) is within 10-3 time steps, the reconstruction error is within 0.005 mm and the accuracy of the virtual stereo-vision system is between the BSS's synchronization error of 10-7 and 10-6 time steps. In addition, after image super-resolution reconstruction technology is applied, the reconstruction error will be reduced to within 0.002 mm. The simulation method proposed in this study can provide a novel research path for existing researchers in the field while also offering the opportunity for researchers without access to costly hardware to participate in related research.

RESUMEN

Detail preservation is a major challenge for single image super-resolution (SISR). Many deep learning-based SISR methods focus on lightweight network design, but these may fall short in real-world scenarios where performance is prioritized over network size. To address these problems, we propose a novel plug-and-play attention module, rich elastic mixed attention (REMA), for SISR. REMA comprises the rich spatial attention module (RSAM) and the rich channel attention module (RCAM), both built on Rich Structure. Based on the results of our research on the module's structure, size, performance, and compatibility, Rich Structure is proposed to enhance REMA's adaptability to varying input complexities and task requirements. RSAM learns the mutual dependencies of multiple LR-HR pairs and multi-scale features, while RCAM accentuates key features through interactive learning, effectively addressing detail loss. Extensive experiments demonstrate that REMA significantly improves performance and compatibility in SR networks compared to other attention modules. The REMA-based SR network (REMA-SRNet) outperforms comparative algorithms in both visual effects and objective evaluation quality. Additionally, we find that module compatibility correlates with cardinality and in-branch feature bandwidth, and that networks with high effective parameter counts exhibit enhanced robustness across various datasets and scale factors in SISR.

RESUMEN

The accurate detection of electrical equipment states and faults is crucial for the reliable operation of such equipment and for maintaining the health of the overall power system. The state of power equipment can be effectively monitored through deep learning-based visual inspection methods, which provide essential information for diagnosing and predicting equipment failures. However, there are significant challenges: on the one hand, electrical equipment typically operates in complex environments, thus resulting in captured images that contain environmental noise, which significantly reduces the accuracy of state recognition based on visual perception. This, in turn, affects the comprehensiveness of the power system's situational awareness. On the other hand, visual perception is limited to obtaining the appearance characteristics of the equipment. The lack of logical reasoning makes it difficult for purely visual analysis to conduct a deeper analysis and diagnosis of the complex equipment state. Therefore, to address these two issues, we first designed an image super-resolution reconstruction method based on the Generative Adversarial Network (GAN) to filter environmental noise. Then, the pixel information is analyzed using a deep learning-based method to obtain the spatial feature of the equipment. Finally, by constructing the logic diagram for electrical equipment clusters, we propose an interpretable fault diagnosis method that integrates the spatial features and temporal states of the electrical equipment. To verify the effectiveness of the proposed algorithm, extensive experiments are conducted on six datasets. The results demonstrate that the proposed method can achieve high accuracy in diagnosing electrical equipment faults.

RESUMEN

The evolution of super-resolution (SR) technology has seen significant advancements through the adoption of deep learning methods. However, the deployment of such models by resource-constrained devices necessitates models that not only perform efficiently, but also conserve computational resources. Binary neural networks (BNNs) offer a promising solution by minimizing the data precision to binary levels, thus reducing the computational complexity and memory requirements. However, for BNNs, an effective architecture is essential due to their inherent limitations in representing information. Designing such architectures traditionally requires extensive computational resources and time. With the advancement in neural architecture search (NAS), differentiable NAS has emerged as an attractive solution for efficiently crafting network structures. In this paper, we introduce a novel and efficient binary network search method tailored for image super-resolution tasks. We adapt the search space specifically for super resolution to ensure it is optimally suited for the requirements of such tasks. Furthermore, we incorporate Libra Parameter Binarization (Libra-PB) to maximize information retention during forward propagation. Our experimental results demonstrate that the network structures generated by our method require only a third of the parameters, compared to conventional methods, and yet deliver comparable performance.

RESUMEN

Recently, ConvNeXt and blueprint separable convolution (BSConv) constructed from standard ConvNet modules have demonstrated competitive performance in advanced computer vision tasks. This paper proposes an efficient model (BCRN) based on BSConv and the ConvNeXt residual structure for single image super-resolution, which achieves superior performance with very low parametric numbers. Specifically, the residual block (BCB) of the BCRN utilizes the ConvNeXt residual structure and BSConv to significantly reduce the number of parameters. Within the residual block, enhanced spatial attention and contrast-aware channel attention modules are simultaneously introduced to prioritize valuable features within the network. Multiple residual blocks are then stacked to form the backbone network, with Dense connections utilized between them to enhance feature utilization. Our model boasts extremely low parameters compared to other state-of-the-art lightweight models, while experimental results on benchmark datasets demonstrate its excellent performance. The code will be available at https://github.com/kptx666/BCRN .

RESUMEN

With the development of deep learning, the Super-Resolution (SR) reconstruction of microscopic images has improved significantly. However, the scarcity of microscopic images for training, the underutilization of hierarchical features in original Low-Resolution (LR) images, and the high-frequency noise unrelated with the image structure generated during the reconstruction process are still challenges in the Single Image Super-Resolution (SISR) field. Faced with these issues, we first collected sufficient microscopic images through Motic, a company engaged in the design and production of optical and digital microscopes, to establish a dataset. Secondly, we proposed a Residual Dense Attention Generative Adversarial Network (RDAGAN). The network comprises a generator, an image discriminator, and a feature discriminator. The generator includes a Residual Dense Block (RDB) and a Convolutional Block Attention Module (CBAM), focusing on extracting the hierarchical features of the original LR image. Simultaneously, the added feature discriminator enables the network to generate high-frequency features pertinent to the image's structure. Finally, we conducted experimental analysis and compared our model with six classic models. Compared with the best model, our model improved PSNR and SSIM by about 1.5 dB and 0.2, respectively.

RESUMEN

Image super-resolution (ISR) is designed to recover lost detail information from low-resolution images, resulting in high-quality and high-definition high-resolution images. In the existing single ISR (SISR) methods based on convolutional neural networks (CNN), however, most of the models cannot effectively combine global and local information and are also easy to ignore the correlation between different hierarchical feature information. To address these problems, this study proposes a multi-level feature interactive image super-resolution network, which is constructed by the convolutional units inspired by nonlinear spiking mechanism in nonlinear spiking neural P systems, including shallow feature processing, deep feature extraction and fusion, and reconstruction modules. The different omni domain self-attention blocks are introduced to extract global information in the deep feature extraction and fusion stage and formed a feature enhancement module having a Transformer structure using a novel convolutional unit for extracting local information. Furthermore, to adaptively fuse features between different hierarchies, we design a multi-level feature fusion module, which not only can adaptively fuse features between different hierarchies, but also can better interact with contextual information. The proposed model is compared with 16 state-of-the-art or baseline models on five benchmark datasets. The experimental results show that the proposed model not only achieves good reconstruction performance, but also strikes a good balance between model parameters and performance.

Asunto(s)

RESUMEN

BACKGROUND AND OBJECTIVE: Magnetic resonance imaging (MRI) can provide rich and detailed high-contrast information of soft tissues, while the scanning of MRI is time-consuming. To accelerate MR imaging, a variety of Transformer-based single image super-resolution methods are proposed in recent years, achieving promising results thanks to their superior capability of capturing long-range dependencies. Nevertheless, most existing works prioritize the design of transformer attention blocks to capture global information. The local high-frequency details, which are pivotal to faithful MRI restoration, are unfortunately neglected. METHODS: In this work, we propose a high-frequency enhanced learning scheme to effectively improve the awareness of high frequency information in current Transformer-based MRI single image super-resolution methods. Specifically, we present two entirely plug-and-play modules designed to equip Transformer-based networks with the ability to recover high-frequency details from dual spaces: 1) in the feature space, we design a high-frequency block (Hi-Fe block) paralleled with Transformer-based attention layers to extract rich high-frequency features; while 2) in the image intensity space, we tailor a high-frequency amplification module (HFA) to further refine the high-frequency details. By fully exploiting the merits of the two modules, our framework can recover abundant and diverse high-frequency information, rendering faithful MRI super-resolved results with fine details. RESULTS: We integrated our modules with six Transformer-based models and conducted experiments across three datasets. The results indicate that our plug-and-play modules can enhance the super-resolution performance of all foundational models to varying degrees, surpassing the capabilities of existing state-of-the-art single image super-resolution networks. CONCLUSION: Comprehensive comparison of super-resolution images and high-frequency maps from various methods, clearly demonstrating that our module possesses the capability to restore high-frequency information, showing huge potential in clinical practice for accelerated MRI reconstruction.

Asunto(s)

RESUMEN

There are several image inverse tasks, such as inpainting or super-resolution, which can be solved using deep internal learning, a paradigm that involves employing deep neural networks to find a solution by learning from the sample itself rather than a dataset. For example, Deep Image Prior is a technique based on fitting a convolutional neural network to output the known parts of the image (such as non-inpainted regions or a low-resolution version of the image). However, this approach is not well adjusted for samples composed of multiple modalities. In some domains, such as satellite image processing, accommodating multi-modal representations could be beneficial or even essential. In this work, Multi-Modal Convolutional Parameterisation Network (MCPN) is proposed, where a convolutional neural network approximates shared information between multiple modes by combining a core shared network with modality-specific head networks. The results demonstrate that these approaches can significantly outperform the single-mode adoption of a convolutional parameterisation network on guided image inverse problems of inpainting and super-resolution.

RESUMEN

Terahertz (THz) waves are electromagnetic waves in the 0.1 to 10 THz frequency range, and THz imaging is utilized in a range of applications, including security inspections, biomedical fields, and the non-destructive examination of materials. However, THz images have a low resolution due to the long wavelength of THz waves. Therefore, improving the resolution of THz images is a current hot research topic. We propose a novel network architecture called J-Net, which is an improved version of U-Net, to achieve THz image super-resolution. It employs simple baseline blocks which can extract low-resolution (LR) image features and learn the mapping of LR images to high-resolution (HR) images efficiently. All training was conducted using the DIV2K+Flickr2K dataset, and we employed the peak signal-to-noise ratio (PSNR) for quantitative comparison. In our comparisons with other THz image super-resolution methods, J-Net achieved a PSNR of 32.52 dB, surpassing other techniques by more than 1 dB. J-Net also demonstrates superior performance on real THz images compared to other methods. Experiments show that the proposed J-Net achieves a better PSNR and visual improvement compared with other THz image super-resolution methods.

RESUMEN

Low-dose computed tomography (LDCT) has been widely used in medical diagnosis. In practice, doctors often zoom in on LDCT slices for clearer lesions and issues, while, a simple zooming operation fails to suppress low-dose artifacts, leading to distorted details. Therefore, numerous LDCT super-resolution (SR) methods have been proposed to promote the quality of zooming without the increase of the dose in CT scanning. However, there are still some drawbacks that need to be addressed in existing methods. First, the region of interest (ROI) is not emphasized due to the lack of guidance in the reconstruction process. Second, the convolutional blocks extracting fix-resolution features fail to concentrate on the essential multi-scale features. Third, a single SR head cannot suppress the residual artifacts. To address these issues, we propose an LDCT CT joint SR and denoising reconstruction network. Our proposed network consists of global dual-guidance attention fusion modules (GDAFMs) and multi-scale anastomosis blocks (MABs). The GDAFM directs the network to focus on ROI by fusing the extra mask guidance and average CT image guidance, while the MAB introduces hierarchical features through anastomosis connections to leverage multi-scale features and promote the feature representation ability. To suppress radial residual artifacts, we optimize our network using the feedback feature distillation mechanism (FFDM) which shares the backbone to learn features corresponding to the denoising task. We apply the proposed method to the 3D-IRCADB and PANCREAS datasets to evaluate its ability on LDCT image SR reconstruction. The experimental results compared with state-of-the-art methods illustrate the superiority of our approach with respect to peak signal-to-noise (PSNR), structural similarity (SSIM), and qualitative observations. Our proposed LDCT joint SR and denoising reconstruction network has been extensively evaluated through ablation, quantitative, and qualitative experiments. The results demonstrate that our method can recover noise-free and detail-sharp images, resulting in better reconstruction results. Code is available at https://github.com/neu-szy/ldct_sr_dn_w_ffdm .

Asunto(s)

RESUMEN

Retinal fundus imaging is a crucial diagnostic tool in ophthalmology, enabling the early detection and monitoring of various ocular diseases. However, capturing high-resolution fundus images often presents challenges due to factors such as defocusing and diffraction in the digital imaging process, limited shutter speed, sensor unit density, and random noise in the image sensor or during image transmission. Super-resolution techniques offer a promising solution to overcome these limitations and enhance the visual details in retinal fundus images. Since the retina has rich texture details, the super-resolution images often introduce artifacts into texture details and lose some fine retinal vessel structures. To improve the perceptual quality of the retinal fundus image, a generative adversarial network that consists of a generator and a discriminator is proposed. The proposed generator mainly comprises 23 multi-scale feature extraction blocks, an image segmentation network, and 23 residual-in-residual dense blocks. These components are employed to extract features at different scales, acquire the retinal vessel grayscale image, and extract retinal vascular features, respectively. The generator has two branches that are mainly responsible for extracting global features and vascular features, respectively. The extracted features from the two branches are fused to better restore the super-resolution image. The proposed generator can restore more details and more accurate fine vessel structures in retinal images. The improved discriminator is proposed by introducing our designed attention modules to help the generator yield clearer super-resolution images. Additionally, an artifact loss function is also introduced to enhance the generative adversarial network, enabling more accurate measurement of the disparity between the high-resolution image and the restored image. Experimental results show that the generated images obtained by our proposed method have a better perceptual quality than the state-of-the-art image super-resolution methods.

Asunto(s)

RESUMEN

Traditional Convolutional Neural Network (ConvNet, CNN)-based image super-resolution (SR) methods have lower computation costs, making them more friendly for real-world scenarios. However, they suffer from lower performance. On the contrary, Vision Transformer (ViT)-based SR methods have achieved impressive performance recently, but these methods often suffer from high computation costs and model storage overhead, making them hard to meet the requirements in practical application scenarios. In practical scenarios, an SR model should reconstruct an image with high quality and fast inference. To handle this issue, we propose a novel CNN-based Efficient Residual ConvNet enhanced with structural Re-parameterization (RepECN) for a better trade-off between performance and efficiency. A stage-to-block hierarchical architecture design paradigm inspired by ViT is utilized to keep the state-of-the-art performance, while the efficiency is ensured by abandoning the time-consuming Multi-Head Self-Attention (MHSA) and by re-designing the block-level modules based on CNN. Specifically, RepECN consists of three structural modules: a shallow feature extraction module, a deep feature extraction, and an image reconstruction module. The deep feature extraction module comprises multiple ConvNet Stages (CNS), each containing 6 Re-Parameterization ConvNet Blocks (RepCNB), a head layer, and a residual connection. The RepCNB utilizes larger kernel convolutions rather than MHSA to enhance the capability of learning long-range dependence. In the image reconstruction module, an upsampling module consisting of nearest-neighbor interpolation and pixel attention is deployed to reduce parameters and maintain reconstruction performance, while bicubic interpolation on another branch allows the backbone network to focus on learning high-frequency information. The extensive experimental results on multiple public benchmarks show that our RepECN can achieve 2.5∼5× faster inference than the state-of-the-art ViT-based SR model with better or competitive super-resolving performance, indicating that our RepECN can reconstruct high-quality images with fast inference.

RESUMEN

In the advancement of medical image super-resolution (SR), the Deep Residual Feature Distillation Channel Attention Network (DRFDCAN) marks a significant step forward. This work presents DRFDCAN, a model that innovates traditional SR approaches by introducing a channel attention block that is tailored for high-frequency features-crucial for the nuanced details in medical diagnostics-while streamlining the network structure for enhanced computational efficiency. DRFDCAN's architecture adopts a residual-within-residual design to facilitate faster inference and reduce memory demands without compromising the integrity of the image reconstruction. This design strategy, combined with an innovative feature extraction method that emphasizes the utility of the initial layer features, allows for improved image clarity and is particularly effective in optimizing the peak signal-to-noise ratio (PSNR). The proposed work redefines efficiency in SR models, outperforming established frameworks like RFDN by improving model compactness and accelerating inference. The meticulous crafting of a feature extractor that effectively captures edge and texture information exemplifies the model's capacity to render detailed images, necessary for accurate medical analysis. The implications of this study are two-fold: it presents a viable solution for deploying SR technology in real-time medical applications, and it sets a precedent for future models that address the delicate balance between computational efficiency and high-fidelity image reconstruction. This balance is paramount in medical applications where the clarity of images can significantly influence diagnostic outcomes. The DRFDCAN model thus stands as a transformative contribution to the field of medical image super-resolution.