Edge-preserving Image Research Articles

Edge enhancement and feature modulation based network for light field depth estimation

Estimating depth from light field images is a critical issue in light field applications. While learning-based methods have made significant strides in light field depth estimation, achieving high accuracy and speed simultaneously remains a major challenge. This paper proposes a light field depth estimation network based on edge enhancement and feature modulation, which significantly improves depth estimation results by emphasizing inter-view correlations while preserving image edge features. Specifically, to prioritize edge details, we introduce an Edge-Enhanced Cost Constructor (EECC) that integrates edge information with existing cost constructors to improve depth estimation performance in complex areas. Furthermore, most light field depth estimation networks utilize only sub-aperture images (SAIs) without considering the inherent angular information in macro-pixel image (MacPI). To address this limitation, we propose the MacPI-Guided Feature Modulation (MGFM) module, which leverages angular information between different views in MacPI to modulate features at each view. Experimental results show that our method not only performs excellently on synthetic datasets but also demonstrates outstanding generalization on real-world datasets, achieving a better balance between accuracy and computation speed.

A multi‐path residual network for image denoising based on edge prior information

AbstractEdge‐preserving denoising is important in image analysis. However, most existing methods suffer from smoothing and loss of high frequency detail at the edges. Aiming at this issue, a multi‐path residual network using edge prior information is proposed. Specifically, the edge is first recovered by a newly designed parallel edge extraction module. Then it is used as a priori to generate affine transform parameters to guide the weight distribution between noise and edge, so that the shallow feature extraction pays more attention to preserving edges. In the deep feature extraction stage, a multi‐scale attention unit is designed to provide the spatial neighbourhood information of the feature points to filter and activate the deep features, which further enhances the ability of the network to extract fine‐grained noisy features. Finally, by taking the difference between the noisy image and the noise feature obtained with residual learning, the denoised image is produced. Extensive experiments show that the PSNR is improved by 0.00–0.78 dB on grey image denoising and 0.00–2.69 dB on colour image denoising compared with others. The denoised image has a better ability to preserve image edges, high frequency details and visual perception.

Gaussian error loss function for image smoothing

Edge-preserving image smoothing plays an important role in the fields of image processing and computational photography, and is widely used for a variety of applications. The edge-preserving filters based on global optimization models have attracted widespread attention due to their nice smoothing quality. According to existing research, the edge-preserving capability is strongly correlated to the penalty function used for gradient regularization. By analyzing the edge-stopping function of existing penalties, we demonstrate that existing image smoothing models are not adequately edge-preserving. In this paper, based on a Gaussian error function (ERF), we propose a Gaussian error loss function (ERLF), which shows stronger edge-preserving capability. We embed the proposed loss function into a global optimization model for edge-preserving image smoothing. In addition, we propose an efficient solution based on additive half-quadratic minimization and Fourier-domain optimization that is capable of processing 720P color images (over 20 fps) in real-time on an NVIDIA RTX 3070 GPU. We have experimented with the proposed filter on a number of low-level vision tasks. Both quantitative and qualitative experimental results show that the proposed filter outperforms existing filters. Therefore, it can be practical for real applications.

Multi exposure fusion for high dynamic range imaging via multi-channel gradient tensor

Multi-exposure fusion (MEF) is an effective technique for directly fusing a sequence of low dynamic range (LDR) images from a high dynamic range (HDR) natural scene. The goal is to generate an information enriched LDR image. Despite its effectiveness, current MEF methods often encounter issues such as detail loss and color degradation. Additionally, existing algorithms often struggle to balance image quality and computation time, particularly for large-sized images. This paper introduces an innovative MEF algorithm that address these challenges, offering improved performance and computational time across all image sizes. The algorithm employs a multi-channel gradient tensor on RGB images to effectively capture the contrast information among the three channels. This mechanism allows an edge-preserving image filter to maintain edges while smoothing weight maps. To enhance computational efficiency, the algorithm uses a fast approximation method suitable for large sized images. Our comprehensive experimental results demonstrate that the proposed method outperforms existing MEF techniques both quantitatively and qualitatively. Furthermore, our method reduces computational time by approximately 30% compared to the most recent state-of-the-art techniques.

Adaptive orthogonal directional total variation with kernel regression for CT image denoising.

Low-dose computed tomography (CT) has been successful in reducing radiation exposure for patients. However, the use of reconstructions from sparse angle sampling in low-dose CT often leads to severe streak artifacts in the reconstructed images. In order to address this issue and preserve image edge details, this study proposes an adaptive orthogonal directional total variation method with kernel regression. The CT reconstructed images are initially processed through kernel regression to obtain the N-term Taylor series, which serves as a local representation of the regression function. By expanding the series to the second order, we obtain the desired estimate of the regression function and localized information on the first and second derivatives. To mitigate the noise impact on these derivatives, kernel regression is performed again to update the first and second derivatives. Subsequently, the original reconstructed image, its local approximation, and the updated derivatives are summed using a weighting scheme to derive the image used for calculating orientation information. For further removal of stripe artifacts, the study introduces the adaptive orthogonal directional total variation (AODTV) method, which denoises along both the edge direction and the normal direction, guided by the previously obtained orientation. Both simulation and real experiments have obtained good results. The results of two real experiments show that the proposed method has obtained PSNR values of 34.5408 dB and 29.4634 dB, which are 1.2392-5.9333 dB and 2.828-6.7995 dB higher than the contrast denoising algorithm, respectively, indicating that the proposed method has good denoising performance. The study demonstrates the effectiveness of the method in eliminating strip artifacts and preserving the fine details of the images.

SAR Image Registration: The Combination of Nonlinear Diffusion Filtering, Hessian Features and Edge Points.

Synthetic aperture radar (SAR) image registration is an important process in many applications, such as image stitching and remote sensing surveillance. The registration accuracy is commonly affected by the presence of speckle noise in SAR images. When speckle noise is intense, the number of image features acquired by single-feature-based methods is insufficient. An SAR image registration method that combines nonlinear diffusion filtering, Hessian features and edge points is proposed in this paper to reduce speckle noise and obtain more image features. The proposed method uses the infinite symmetric exponential filter (ISEF) for image pre-processing and nonlinear diffusion filtering for scale-space construction. These measures can remove speckle noise from SAR images while preserving image edges. Hessian features and edge points are also employed as image features to optimize the utilization of feature information. Experiments with different noise levels, geometric transformations and image scenes demonstrate that the proposed method effectively improves the accuracy of SAR image registration compared with the SIFT-OCT, SAR-SIFT, Harris-SIFT, NF-Hessian and KAZE-SAR algorithms.

Image smoothing method based on global gradient sparsity and local relative gradient constraint optimization

Removing texture while preserving the main structure of an image is a challenging task. To address this, this paper propose an image smoothing method based on global gradient sparsity and local relative gradient constraints optimization. To reduce the interference of complex texture details, adopting a multi-directional difference constrained global gradient sparsity decomposition method, which provides a guidance image with weaker texture detail gradients. Meanwhile, using the luminance channel as a reference, edge-aware operator is constructed based on local gradient constraints. This operator weakens the gradients of repetitive and similar texture details, enabling it to obtain more accurate structural information for guiding global optimization of the image. By projecting multi-directional differences onto the horizontal and vertical directions, a mapping from multi-directional differences to bi-directional gradients is achieved. Additionally, to ensure the consistency of measurement results, a multi-directional gradient normalization method is designed. Through experiments, we demonstrate that our method exhibits significant advantages in preserving image edges compared to current advanced smoothing methods.

Deep unsupervised nonconvex optimization for edge-preserving image smoothing

Deep unsupervised nonconvex optimization for edge-preserving image smoothing

Adaptive Bayesian speckle reduction in the 2D DOST domain using 2D MC-GARCH-M model: Preserving image edges and textural information

Image denoising and image quality enhancement are major issues in image processing. Additive and multiplicative noise removal is one of the main image enhancement approaches. In this paper, we propose a novel 2D Merged Complex Generalized Autoregressive Conditional Heteroscedasticity Mixture (MC-GARCH-M) model for 2D complex stochastic processes. This model effectively captures non-Gaussian statistics and all three types of dependencies within the processes (RRD, IID, and RID) using the proposed linear mapping transformation T. Our statistical analysis shows that the 2D MC-GARCH-M model provides the best fit to the non-Gaussian statistics of the mapped matrix of 2D DOST coefficients compared to Gaussian, Generalized Gaussian, and Stable distributions. We confirm the presence of heteroscedasticity in the mapped matrix using the Engle hypothesis test, verifying ARCH/GARCH effects. The 2D MC-GARCH-M model, with its location-dependent conditional variances, provides a flexible approach for noise removal and modeling complex stochastic processes. Experimental results on artificially speckled aerial and actual SAR images demonstrate that our method outperforms state-of-the-art approaches in terms of speckle removal and preserving the edges and textural information of the image. Key innovations include: a new statistical analysis using linear mapping T showing non-Gaussian heavy-tailed distributions; an adaptive Bayesian estimator, MCMAP, based on the heteroscedastic model for speckle noise removal; and the method's applicability to images with varying mutual information (MI) between real and imaginary parts of 2D DOST coefficients. Our method is adaptive and robust to initial parameter settings, effectively utilizing magnitude and phase information, providing an optimal closed-form solution, reducing memory and computational requirements, and demonstrating robustness to speckle power levels. While requiring more computational resources compared to thresholding methods, our approach is less demanding than other Bayesian algorithms, making it suitable for offline processing.

Edge-preserving image deraining network using cumulative feature aggregation

This study proposes an edge-preserving image deraining network using a wavelet feature aggregation method. Wavelet subbands re correlated with each other, and the high-frequency subband in the horizontal direction is the least affected by rain streak contamination. On this basis, we introduce a single image deraining network that cumulatively aggregates wavelet subband features according to their importance. The network architecture primarily comprises a wavelet feature aggregation block and a residue channel guide block. The aggregation of features with the cumulative wavelet feature aggregation block moves downward and upward, and a long short-term memory-based multiscale attentive rain streak removal block is developed to serve as the backbone for rain streak removal. We use a residual channel map based on the low-frequency subband to construct guide features that assist in rain streak removal. A repetitive image restoration framework that incorporates two proposed blocks is used to iteratively improve rainy images. We test the proposed network on various image datasets and compare the deraining performance with those of existing methods The experimental results demonstrate that the performance of the proposed scheme is superior to that of other tested deraining methods. https://github.com/SYChoi98/CWFANet/

Edge-preserving image restoration based on a weighted anisotropic diffusion model

Partial differential equation-based methods have been widely applied in image restoration. The anisotropic diffusion model has a good noise removal capability without affecting significant edges. However, existing anisotropic diffusion-based models closely depend on the diffusion coefficient function and threshold parameter. This paper proposes a new weighted anisotropic diffusion coefficient model with multiple scales, and it has a higher speed of closing to X-axis and exploits adaptive threshold parameters. Meanwhile, the proposed algorithm is verified to be suitable for multiple types of noise. Numerical metrics and visual comparison of simulation experiments show the proposed model has significant superiority in edge-preserving and staircase artifacts reducing over the existing anisotropic diffusion-based techniques.

Hyperbolic tangent penalty function for edge-preserving image filtering

Edge-preserving image filtering provides a common interface for a variety of edge-aware image processing tasks. It is essential to the field of image processing and computational photography. Recent attempts at optimization models with non-convex regularization have shown promising performance in edge-preserving image filtering. In this paper, we propose a novel non-convex regularization term based on a hyperbolic tangent penalty function, which is shown to be more edge-preserving than popular non-convex penalty functions. We embed the proposed regularization term in an optimization model for edge-preserving image filtering. To solve the model with the non-convex regularization, we propose an efficient solution based on the additive half quadratic minimization and Fourier domain optimization. We have conducted extensive experiments to evaluate the proposed filter. Both quantitative and qualitative results demonstrate that our filter benefits various applications. Furthermore, the proposed filter is highly efficient and renders real-time processing of 720P color images on a modern GPU.

Robust edge-preserving image smoothing based on complementary weighting scheme

Robust edge-preserving image smoothing based on complementary weighting scheme

Bilateral regularized optimization model for edge-preserving image smoothing

Edge-preserving image smoothing is vital in the field of image processing and computational photography. The state-of-the-art filters based on optimization models have achieved promising performance. However, most of them fail to consider the spatial support in the regularization term, thus limiting the edge-preserving capabilities. In this paper, inspired by the bilateral filter, which consists of a range kernel and a spatial kernel. we propose to leverage bilateral kernel as a penalty function, and embed it into an optimization model for edge-preserving image smoothing. Furthermore, we propose to incorporate an edge-aware weighted scheme in the data term design, which further improves the edge-preserving capability. The bilateral function is non-convex and can be non-trivial to solve. In this paper, we propose a novel iterative solution based on fixed point iteration, where the main burden in each iteration is a bilateral filtering process. We have conducted extensive experiments to evaluate the proposed filter. Experiment results indicate that our filter benefits a variety of image processing tasks. Moreover, we propose an efficient approximation of the proposed filter, which is able to significantly accelerate the filtering process with neglectable sacrifice of smoothing quality.

An Improved Variational-Based Model for Denoising and Segmentation of Vector-Valued Images

Preserving important features such as edges is one of the main concerns in models for denoising and segmenting vector-valued (colour) images. The Rudin-Osher-Fatemi (ROF) model is a well-known variational-based image denoising model that is capable of reducing image noise while preserving image edges. However, the ROF model is not formulated for denoising colour images and is less effective in preserving corners and weak edges. On the other hand, a variational-based selective segmentation model for colour images called the selective distance segmentation (DSS2) model has recently been proposed, which can effectively partition or extract a specific object in an image. However, the DSS2 model has problems in segmenting colour images with noise, which may result in poor segmentation. Therefore, in this research, we first modify the ROF model to denoise vector-valued images by including the edge detector and extending the formulation into a vector-valued framework. Second, we reformulate the DSS2 model by incorporating the modified ROF model as a new fitting term in the DSS2 model. Peak signal-to-noise ratio (PSNR) is used to measure the image quality, while Jaccard and Dice similarity index are used to evaluate the segmentation quality. The comparison between our proposed model and existing model shows that our model is more effective as indicated by higher PSNR, Jaccard and Dice similarity index values.

Pulsed radiation image restoration based on unsupervised deep learning

The process of pulsed X-ray imaging generates noise and blurring that significantly degrade image quality, lead to loss of image information, and interfere with accurate diagnosis of X-ray spot size and objects. Traditional radiation image restoration methods lack flexibility and generalization. Furthermore, neural network methods based on supervised learning require many actual image data sets to be obtained in advance, which limits network training. In this paper, we propose an unsupervised radiation image restoration method based on a deep image prior. The classical DeepRED framework is extended by adding weight-adaptive variational regularization. Automatic strategies are employed to estimate local regularization parameters, resulting in better noise removal and preservation of image edges and texture structures. The proposed method is evaluated through numerical simulations and experimental studies of X-ray source images and flash radiographic images. The results indicate that the proposed method is effective in restoring pulsed radiation images, removing noise, and preserving boundary region information. Additionally, this method produces satisfactory results for the restoration of real neutron images, demonstrating its applicability in the field of radiation imaging.

Untrained neural network embedded Fourier phase retrieval from few measurements

Fourier phase retrieval (FPR) is a challenging task used in various applications to recover an unknown signal from its Fourier phaseless measurements. FPR with few measurements helps reduce time and hardware costs but suffers from severe ill-posedness. Recently, untrained neural networks have offered new approaches by introducing learned priors to alleviate ill-posedness without requiring external data. However, they are computationally expensive and not ideal for reconstructing high-frequency structures in images. This paper proposes an untrained neural network (NN) embedded FPR algorithm, based on the alternating direction method of multipliers. Specifically, we use a generative network to represent the image to be recovered, which confines the image to the space defined by the network structure. Additionally, total variation regularization is imposed to facilitate the preservation of image edges and local structures. Furthermore, to reduce the computational cost mainly caused by the updates of NN’s parameters, we develop an accelerated algorithm that adaptively trades off between explicit and implicit regularization. We theoretically analyze that the proposed algorithm can converge to a stationary point of the objective function under mild conditions. Experimental results indicate that the proposed algorithm outperforms existing untrained-NN-based algorithms with fewer computational resources and even performs competitively against trained-NN-based algorithms.

Atelectasis detection in chest X-ray images using convolutional neural networks and transfer learning with anisotropic diffusion filter

Atelectasis is the loss of volume caused by decreasing gas in a specific area of the lung. It occurs when the lung sacs (alveoli) do not fully inflate, resulting in a lack of oxygen to the blood, tissues, organs, and fill with alveolar fluid. It can be caused by pressure outside of your lung, an obstruction, poor airflow, or scarring. It is critical to diagnose this lung condition as soon as possible. Chest X-rays are the most commonly utilized diagnostic tool for this condition. Examining chest X-rays, however, is difficult even for a professional radiologist. There is a need to improve diagnosis accuracy. As a result, this study proposes a novel detection and classification approach for rapid diagnosis of atelectasis utilizing patient chest X-ray data. To diagnose atelectasis from chest X-ray images, we used state-of-the-art models like VGG19, Inception, and a deep learning method (CNN). This study presents an effective method for categorizing chest X-ray images as normal or atelectasis-infected. This study offers a convolutional neural network (CNN) method to aid medical experts in identifying atelectasis disease. The anisotropic diffusion filtering (ADF) approach was used to improve image edge preservation, reduce noise, and contrast limited adaptive histogram equalization (CLAHE) for improving the contrast of low-intensity images. After evaluating the CNN model, it achieved 99.88 % training accuracy, 99 % validation accuracy, and 99 % test accuracy. In this study, CNN achieved a result that outperformed state-of-the-art models (VGG19 and Inception). As a result of the findings, deep features provided consistent and reliable features for detecting atelectasis. Therefore, the suggested method expedites atelectasis diagnosis and radiologists' screening of atelectasis patients.

Image edge preservation via low-rank residuals for robust subspace learning

Image edge preservation via low-rank residuals for robust subspace learning

An Enhance Relative Total Variation With BF Model for Edge-Preserving Image Smoothing

In image processing, edge-preserving image smoothing methods that maintain weak structure while smoothing multiscale textures with strong gradients remain challenging. In this paper, a new global optimization method named Enhance Relative Total Variation Embedded with Bilateral Filtering is proposed. The texture and structure are intuitively considered separately from the model generalization. First, the weak structures are separated from the over-penalized texture and structure items by embedding bilateral filtering. Then, the in-set shrinking edges/structures are gradually recovered by constructing a contrast stretching function. In comparison to current state-of-the-art methods, experimental results demonstrate that the method is effective in maintaining weak structures and suppressing multiscale textures.