SED-Net: Single-input multiple-output network for simultaneous low-light image enhancement and deblurring

Single blind image deblurring: advances and prospects

D2LA-Net: optimized dual-decoder architecture for simultaneous image deblurring and depth estimation

During spacecraft launch, flight, and docking operations, monitored images often suffer from spatially variant blur caused by atmospheric turbulence, defocusing, and relative motion. To address this challenge, we propose a novel unsupervised deblurring framework tailored specifically for spacecraft imagery. Our approach incorporates three key innovations: First, we design a detail-preserving local region selection strategy based on multi-scale morphological gradients with adaptive thresholding, which optimizes regions for blur kernel estimation. Second, we define a blur kernel error term and integrate it into the degradation model, introducing explicit error correction constraints into the alternating iterative minimization process. Third, we incorporate Shearlet transform regularization to enhance the recovery of fine local details. Experimental results demonstrate that our method significantly outperforms state-of-the-art unsupervised techniques and even surpasses several advanced deep learning approaches in preserving complex structural details under spatially variant degradation. Our code and data are available at https://github.com/bsfsf/Image_deblur.

Total curvature-driven blind image deblurring

DFFormer: Dual Frequency-Driven Transformer for real-world image deblurring

Wide-field nailfold capillary image deblurring method based on an improved MIMO-UNet.

This study introduces a novel inertial-type iteration algorithm based on the Normal S iteration for the class of almost contraction mappings in Banach spaces. Traditional fixed point iterations often suffer from slow convergence and high computational cost; to address these limitations, the proposed framework incorporates an adaptive inertial-type parameter. We establish strong convergence of the algorithm and derive explicit a posteriori error estimates under weak contractive conditions. In addition, we demonstrate the asymptotic equivalence of the NS inertial-type trajectories with the classical Normal S iteration, provide a comprehensive weak w2−stability analysis, and obtain sharp upper bounds for the data dependence problem. The practical performance of the algorithm is evaluated in two distinct computational domains: image deblurring via wavelet-based ℓ1 regularization and the generation of complex fractal patterns, including Julia and Mandelbrot sets. Numerical results show that the proposed inertial-type iteration algorithm significantly outperforms existing methods—such as Picard, Mann, Ishikawa, and standard Normal S iterations—achieving faster convergence, higher PSNR values in image restoration, and more stable basins of attraction in fractal visualizations. These findings highlight the effectiveness and versatility of the NS inertial-type iteration algorithm approach for both theoretical analysis and real-world applications.

In uncrewed aerial vehicle (UAV)-based ground observation and detection missions involving high-speed moving targets or low-light conditions, Time Delay Integration (TDI) cameras enhance image brightness through multi-stage charge accumulation. However, the imaging quality is susceptible to motion blur induced by platform vibrations and velocity mismatch. Based on TDI imaging technology, a TDI image degradation model for a UAV-based imaging platform is formulated. To address spatial blurring caused by platform vibration and velocity mismatch during TDI imaging, we propose a TDI image restoration algorithm based on sharpness-optimized partitioned Point Spread Function (PSF) estimation. The main innovation lies in the first application of partitioned PSF estimation combined with image sharpness optimization in TDI imaging. By formulating an accurate TDI image degradation model, spatial motion blur kernel estimation is transformed into an iterative search problem for partitioned optimal PSF. Solving for optimal sharpness yields the optimal PSF and corresponding local motion parameters, achieving image restoration. Simulation and experimental results demonstrate that the proposed algorithm in this paper effectively removes motion blur in TDI dynamic imaging, while suppressing artifacts and ringing, thus significantly enhancing image quality.

Image restoration is a vital research area in computer vision, focusing on reconstructing high-quality clear images from degraded observations. Common types of degradation include noise and blur, which may stem from imaging device limitations, environmental interference, and other factors. This paper centers on the design and optimization of multi-stage image restoration networks, conducting in-depth exploration of feature extraction, feature fusion, attention mechanisms, and their practical applications. A multi-stage hybrid attention mechanism-based image restoration network is proposed. Initially, each stage progressively extracts and restores image features. Then, an adaptive feature fusion block enables effective cross-stage information transfer. Finally, by calculating losses at each stage and assigning different weights, the network achieves stable convergence during training. The hybrid attention mechanism enhances the model’s focus on critical features and improves its understanding of the overall image structure. Outstanding performance has been achieved in both image deblurring and denoising tasks. On the GoPro dataset, the restored results achieved a PSNR of 33.26 and an SSIM of 0.963. On the SIDD dataset, the restored results reached a PSNR of 40.23 and an SSIM of 0.963. Furthermore, ablation experiments demonstrated the effectiveness of the multi-stage model, hybrid attention mechanism, and adaptive feature fusion block.

An efficient and lightweight pyramid attention for image deblurring

To address the issues of insufficient restoration of texture details in deblurred images and inadequate learning of frequency domain features, an image deblurring algorithm based on frequency domain feature enhancement and convolutional neural networks is proposed. In this architecture, firstly, a Fourier residual module with a parallel structure is constructed to achieve collaborative learning and modeling of spatial and frequency domain features, aiming to improve frequency domain feature learning capability and the restoration effect of the texture details; secondly, a gated controlled feed-forward unit acts on the Fourier residual module to further enhance the nonlinear expression ability of the algorithm; thirdly, a supervised attention module is improved and added to the decoder to promote more effective capture of key features for image reconstruction; finally, the weighted sum of spatial domain Charbonnier loss function and frequency domain loss function is defined as a novel total loss function. In addition, to verify the performance of our proposed algorithm, we conducted experiments on the GOPRO and HIDE datasets. Through experiments on the GOPRO, we obtained an SSIM and an LPIPS of 0.961 and 0.0278, respectively. With regard to the experiments on the HIDE datasets, we obtained an SSIM and an LPIPS of 0.941 and 0.0286, respectively. As for parameter count and running time, their values were 1.197 and 9.15 × 106, respectively, obtained by the experiments on the GOPRO. In all algorithms, the values of our proposed algorithm are optimal. However, the PSNR of our proposed algorithm is very close to that of the latest comparison algorithm and is suboptimal. In a word, experimental results have demonstrated that our proposed algorithm effectively removes blur while better preserving the details and edges of the image. Therefore, it has more practical value and prospects in computer vision tasks.

Transformers have excelled in image restoration due to their advanced representational abilities. However, their reliance on a fixed local window for attention often undermines translation invariance and local relationship preservation. This limitation can reduce network stability, especially when dealing with positional changes in degradation scenarios. In this research, we present a new Bayesian Window Transformer, which innovates by employing a probability distribution for window shifts, overcoming the limitations of fixed window configurations in traditional transformers. This approach allows for more flexible coverage beyond a predetermined region. During the evaluation procedure, we further develop two approximate inference algorithms: Layer Expectation Propagation and Monte Carlo Average. These two algorithms calculate expectations derived from the introduced distribution to effectively approximate the marginalization results of the probabilistic variables. Hence, our Bayesian Window Transformer not only inherits the powerful representation ability but also maintains essential properties like translation invariance and local relationship preservation for image restoration. We also provide a theoretical guarantee, demonstrating that our method is aligned with the classic sliding window technique in terms of receptive field sizes and sliding behavior. Comprehensive experiments validate the exceptional effectiveness of our Bayesian Window Transformer across multiple image restoration tasks, including image deraining, denoising, and deblurring.

Exploring the transformer-based and diffusion-based models for single image deblurring

ABSTRACT In this paper, we mainly investigate the backward problem for the two‐dimensional time‐fractional diffusion equation by the proper orthogonal decomposition (POD) method, and also apply this method to solve the image deblurring problem. The approach constructs a low‐dimensional basis by solution snapshots to accelerate the reconstruction of the initial field from noisy terminal time measurements. Besides, we give the Tikhonov regularization method and obtain the convergence analysis about the regularization solutions. Finally, we propose several numerical experiments to show the efficiency and accuracy of our method.

ABSTRACT Accurate restoration of remote sensing imagery is challenged by complex degradation factors, including atmospheric turbulence and sensor imperfections. A major bottleneck lies in the difficulty of disentangling atmospheric effects from the system’s intrinsic Point Spread Function (PSF), which limits the stability and generalization of conventional restoration methods. This study proposes MoonPSF-Decon, which derives an intrinsic, sensor-consistent PSF from high-contrast, atmosphere-free FY-4B/AGRI lunar observations and applies it to Earth imagery for non-blind deconvolution using Wiener (WN) and Lucy – Richardson (LR) deconvolution algorithms. Unlike prior lunar-target studies focused on Modulation Transfer Function (MTF) diagnostics, we demonstrate short-term PSF reusability: a single MoonPSF extracted on 28 September 2023 is reused to restore a month-long sequence of AGRI Earth images (1–31 October 2023), improving operational efficiency. A progressive validation strategy – including simulation, real-scene restoration, and cross-sensor comparison with near-synchronous Himawari-9/AHI high-resolution data – confirms the method’s effectiveness. For AGRI 10.8 µm imagery, peak signal-to-noise ratio (PSNR) increases by 1.8% for both LR and WN, root mean square error (RMSE) decreases by 8.1% (LR) and 8.0% (WN), and structural similarity (SSIM) improves by 0.75% (LR) and 0.8% (WN). The month-long evaluation shows consistently improved PSNR/SSIM and reduced RMSE relative to the AHI quasi-reference, supporting the temporal robustness of the lunar-derived intrinsic PSF. Overall, MoonPSF-Decon provides a stable and physically grounded PSF prior for practical, multi-temporal Earth observation image deblurring.

Ultra-High-Definition (UHD) image restoration has acquired remarkable attention due to its practical demand. In this paper, we construct UHD snow and rain benchmarks, named UHD-Snow and UHD-Rain, to remedy the deficiency in this field. The UHD-Snow/UHD-Rain is established by simulating the physics process of rain/snow into consideration and each benchmark contains 3200 degraded/clear image pairs of 4K resolution. Furthermore, we propose an effective UHD image restoration solution by considering gradient and normal priors in model design, thanks to these priors’ spatial and detail contributions. Specifically, our method contains two branches: (a) feature fusion and reconstruction branch in high-resolution space and (b) prior feature interaction branch in low-resolution space. The former learns high-resolution features and fuses prior-guided low-resolution features to reconstruct clear images, while the latter utilizes normal and gradient priors to mine useful spatial features and detail features to guide high-resolution recovery better. To better utilize these priors, we introduce single prior feature interaction and dual prior feature interaction, where the former respectively fuses normal and gradient priors with high-resolution features to enhance prior ones, while the latter calculates the similarity between enhanced prior ones and further exploits dual guided filtering to boost the feature interaction of dual priors. We conduct experiments on both new and existing public datasets and demonstrate the state-of-the-art performance of our method on UHD image low-light enhancement, dehazing, deblurring, desnowing, and deraining. The source codes and benchmarks are available at https://github.com/wlydlut/UHDDIP.

A multi-scale gate network for high-quality image deblurring

Weighted Patch-wise Minimum Pixel Prior for Blind Image Deblurring

MSTSGM: A multi-scale temporal–spatial guided model for image deblurring