Edge-preserving Image Research Articles

A l1 Norm Based Image Prior Combination in Multiframe Superresolution

We address the multiframe superresolution problem using the variational Bayesian method in this paper. In the variational Bayesian framework, the prior is crucial in transferring the ill-posed reconstruction problem to a well-posed one. We propose a prior combination method based on filter bank and l1 norm. Multiple filters are used in our prior model, and the corresponding combination coefficient vector can be estimated by the characteristics of the filtered image and noise. Furthermore, the local adaptive coefficients of every filter are more effective in removing noise and preserving image edges. Extensive experiments demonstrate the advantages of the proposed method.

Effective Strip Noise Removal for Low-Textured Infrared Images Based on 1-D Guided Filtering

Infrared images typically contain obvious strip noise. It is a challenging task to eliminate such noise without blurring fine image details in low-textured infrared images. In this paper, we introduce an effective single-image-based algorithm to accurately remove strip-type noise present in infrared images without causing blurring effects. First, a 1-D row guided filter is applied to perform edge-preserving image smoothing in the horizontal direction. The extracted high-frequency image part contains both strip noise and a significant amount of image details. Through a thermal calibration experiment, we discover that a local linear relationship exists between infrared data and strip noise of pixels within a column. Based on the derived strip noise behavioral model, strip noise components are accurately decomposed from the extracted high-frequency signals by applying a 1-D column guided filter. Finally, the estimated noise terms are subtracted from the raw infrared images to remove strips without blurring image details. The performance of the proposed technique is thoroughly investigated and is compared with the state-of-the-art 1-D and 2-D denoising algorithms using captured infrared images.

Superpixel-Guided Adaptive Image Smoothing

In edge-preserving image smoothing, edge blurriness and structural edge attenuation have been common problems. $L_0$ smoothing successfully solves these two problems by adopting $L_0$ norm of gradients. However, a weak structural edge diminishing problem still exists because $L_0$ penalty first removes small nonzero gradients. In order to address this problem, we propose superpixel-guided adaptive image smoothing by introducing an adaptive parameter into $L_0$ smoothing framework. The adaptive smoothing parameter is efficiently computed in a cascade manner. In the first stage, we allocate smoothing parameters to the pixels consisting of details. More importantly, we then exploit similarities between a pixel and its surrounding superpixels for assigning smoothing parameters to the rest pixels. Experimental results demonstrate that our proposed method efficiently preserves structural edges regardless of their scales compared to previous methods.

An effective edge-preserving smoothing method for image manipulation

This paper presents a novel and effective edge-preserving image smoothing method for edge-aware image manipulation. The method formulates the smoothing as a problem of minimizing a convex object function with a constraint and an efficient solution to the optimization problem is presented. Specifically, the method provides an unified framework to regularize the edge and texture pixels in the optimization so that geometric edges representing image structures can be well retained and fine edges of texture regions are removed or suppressed. Both qualitative and quantitative experimental results on natural images and computer-generated structured images have shown the efficacy of the proposed method. In addition, the proposed method can improve the performance of many image processing and manipulation tasks including edge extraction and simplification, non-photorealistic rendering, detail and contrast exaggeration, HDR tone mapping, block-based discrete cosine transform (BDCT) artifact removal and content-aware image resizing, as demonstrated through the experiments.

MO-DE-207A-11: Sparse-View CT Reconstruction Via a Novel Non-Local Means Method

Purpose: Sparse-view computed tomography (CT) reconstruction is an effective strategy to reduce the radiation dose delivered to patients. Due to its insufficiency of measurements, traditional non-local means (NLM) based reconstruction methods often lead to over-smoothness in image edges. To address this problem, an adaptive NLM reconstruction method based on rotational invariance (RIANLM) is proposed. Methods: The method consists of four steps: 1) Initializing parameters; 2) Algebraic reconstruction technique (ART) reconstruction using raw projection data; 3) Positivity constraint of the image reconstructed by ART; 4) Update reconstructed image by using RIANLM filtering. In RIANLM, a novel similarity metric that is rotational invariance is proposed and used to calculate the distance between two patches. In this way, any patch with similar structure but different orientation to the reference patch would win a relatively large weight to avoid over-smoothed image. Moreover, the parameter h in RIANLM which controls the decay of the weights is adaptive to avoid over-smoothness, while it in NLM is not adaptive during the whole reconstruction process. The proposed method is named as ART-RIANLM and validated on Shepp-Logan phantom and clinical projection data. Results: In our experiments, the searching neighborhood size is set to 15 by 15 and the similarity window is set to 3 by 3. For the simulated case with a resolution of 256 by 256 Shepp-Logan phantom, the ART-RIANLM produces higher SNR (35.38dB<24.00dB) and lower MAE (0.0006<0.0023) reconstructed image than ART-NLM. The visual inspection demonstrated that the proposed method could suppress artifacts or noises more effectively and preserve image edges better. Similar results were found for clinical data case. Conclusion: A novel ART-RIANLM method for sparse-view CT reconstruction is presented with superior image. Compared to the conventional ART-NLM method, the SNR and MAE from ART-RIANLM increases 47% and decreases 74%, respectively.

Noise‐insensitive and edge‐preserving resolution upconversion scheme for digital image based on the spatial general autoregressive model

This study proposes an edge-preserving image interpolation algorithm for both clean images and polluted images. First, the structure of an image window is learnt adaptively by using a spatial general autoregressive (SGAR) model that is a uniform expression for both linear and non-linear autoregressive (AR) models. Parameters of the SGAR model are estimated in a moving window in the input low-resolution image by using the robust generalised M-estimator. Next, the interpolation model is established from the learnt model and a new feedback mechanism in accordance with the residual sum of squares minimisation principle. Finally, the gradient simulated annealing algorithm is used to solve the interpolation model, which can rapidly converge to the global optimum in probability with the help of gradient information. Experiments have been performed using worldwide datasets to evaluate the performance of the authors method. The results demonstrate that their method is superior to a recent AR model-based method and is bicubic, especially when images are polluted by noise such as Gaussian noise, Poisson noise, impulse noise, or a combination of these.

Fractional-order total variation combined with sparsifying transforms for compressive sensing sparse image reconstruction

The total variation (TV) model has been considered to be one of the most successful and representative model for compressive sensing (CS) sparse image reconstruction due to its advantage of preserving image edges. However, TV regularized term often favors piecewise constant solution and therefore it fails to preserve the details and textures. To overcome this defect and reconstruct the fine details, this paper proposes a two-dimensional CS sparse image reconstruction model by introducing the fractional-order TV (FrTV) regularization constraint into CS optimization problem. Furthermore, in order to achieve sparser representation flexibly, a combination of discrete wavelet transform and curvelet transform ℓ1-norm regularization is also incorporated into the cost function and a method for estimating the regularization parameter that trades off the two terms in the cost function is proposed. By using a smooth approximation of the ℓ1-norm, a gradient projection algorithm is derived to solve the combined FrTV and sparsifying transforms constrained minimization problem effectively. Compared with several state-of-the-art reconstruction algorithms, the proposed algorithm is more efficient and robust, not only yielding higher peak-signal-to-noise ratio, but also reconstructing the fine details and textures more efficiently.

Improved spatial regression analysis of diffusion tensor imaging for lesion detection during longitudinal progression of multiple sclerosis in individual subjects

Subject-specific longitudinal DTI study is vital for investigation of pathological changes of lesions and disease evolution. Spatial Regression Analysis of Diffusion tensor imaging (SPREAD) is a non-parametric permutation-based statistical framework that combines spatial regression and resampling techniques to achieve effective detection of localized longitudinal diffusion changes within the whole brain at individual level without a priori hypotheses. However, boundary blurring and dislocation limit its sensitivity, especially towards detecting lesions of irregular shapes. In the present study, we propose an improved SPREAD (dubbed improved SPREAD, or iSPREAD) method by incorporating a three-dimensional (3D) nonlinear anisotropic diffusion filtering method, which provides edge-preserving image smoothing through a nonlinear scale space approach. The statistical inference based on iSPREAD was evaluated and compared with the original SPREAD method using both simulated and in vivo human brain data. Results demonstrated that the sensitivity and accuracy of the SPREAD method has been improved substantially by adapting nonlinear anisotropic filtering. iSPREAD identifies subject-specific longitudinal changes in the brain with improved sensitivity, accuracy, and enhanced statistical power, especially when the spatial correlation is heterogeneous among neighboring image pixels in DTI.

Spatiotemporal Rank Filtering Improves Image Quality Compared to Frame Averaging in 2-Photon Laser Scanning Microscopy.

Live imaging of biological specimens using optical microscopy is limited by tradeoffs between spatial and temporal resolution, depth into intact samples, and phototoxicity. Two-photon laser scanning microscopy (2P-LSM), the gold standard for imaging turbid samples in vivo, has conventionally constructed images with sufficient signal-to-noise ratio (SNR) generated by sequential raster scans of the focal plane and temporal integration of the collected signals. Here, we describe spatiotemporal rank filtering, a nonlinear alternative to temporal integration, which makes more efficient use of collected photons by selectively reducing noise in 2P-LSM images during acquisition. This results in much higher SNR while preserving image edges and fine details. Practically, this allows for at least a four fold decrease in collection times, a substantial improvement for time-course imaging in biological systems.

Quadrilateral Scaled Threshold Weighted Averaging Filter for MRI Denoising

Image denoising of Magnetic Resonance Imaging (MRI) by manifold approach focuses on the removal of inhomogeneous Gaussian and Rician noise without blurring the fine anatomical structures. Non Local Means (NLM) filter, the popular denoising technique, does not hold well for edge preservation of images with low Signal to Noise Ratio (SNR). Hence, an alternative formulation of NLM filter is required. This paper presents a quadrilateral scaling method based on image gradients for change in the isotropic nature of NLM and provides an estimate for edge orientation. A new definition of NLM weights based on similarity intensity measures that reflect on the noise statistics is introduced. The efficacy and feasibility of the proposed method is demonstrated through simulation. The probable reduction in time consumption is illustrated.

An adaptive edge-preserving image denoising technique using patch-based weighted-SVD filtering in wavelet domain

Image denoising has always been one of the standard problems in image processing and computer vision. It is always recommendable for a denoising method to preserve important image features, such as...

A self-learning image super-resolution method via sparse representation and non-local similarity

It is difficult to design an image super-resolution algorithm that can not only preserve image edges and texture structure but also keep lower computational complexity. A new super-resolution model based on sparsity regularization in Bayesian framework is presented. The fidelity term restricts the underlying image to be consistent with the observation image in terms of the image degradation model. The sparsity regularization term constraints the underlying image with a sparse representation in a proper dictionary. The non-local self-similarity is also introduced into the model. In order to make the sparse domain better represent the underlying image, we use high-frequency features extracted from the underlying image patches for sparse representation, which increases the effectiveness of sparse modeling. The proposed method learns dictionary directly from the estimated high-resolution image patches (extracted features), and the dictionary learning and the super-resolution can be fused together naturally into one coherent and iterated process. Such a self-learning method has stronger adaptability to different images and reduces dictionary training time. Experiments demonstrate the effectiveness of the proposed method. Compared with some state-of-the-art methods, the proposed method can better preserve image edges and texture details.

An Edge-Preserved Image Denoising Algorithm Based on Local Adaptive Regularization

Image denoising methods are often based on the minimization of an appropriately defined energy function. Many gradient dependent energy functions, such as Potts model and total variation denoising, regard image as piecewise constant function. In these methods, some important information such as edge sharpness and location is well preserved, but some detailed image feature like texture is often compromised in the process of denoising. For this reason, an image denoising method based on local adaptive regularization is proposed in this paper, which can adaptively adjust denoising degree of noisy image by adding spatial variable fidelity term, so as to better preserve fine scale features of image. Experimental results show that the proposed denoising method can achieve state-of-the-art subjective visual effect, and the signal-noise-ratio (SNR) is also objectively improved by 0.3–0.6 dB.

Estimating intravoxel fiber architecture using constrained compressed sensing combined with multitensor adaptive smoothing

In diffusion magnetic resonance imaging dMRI, the accuracy of fiber tracking and analysis depends directly on that of intravoxel fiber architecture reconstruction. Several methods have been proposed that estimate intravoxel fiber architecture using low angular resolution acquisitions owing to their shorter acquisition time and relatively low b-values. But these methods are highly sensitive to noise. We propose an approach to estimating intravoxel fiber architecture in low angular resolution dMRI. The method consists in using a constrained compressed sensing CCS method, also known as crossing fiber angular resolution of intravoxel architecture CFARI technique, in combination with multitensor adaptive smoothing in which a diffusion-weighted DW image smoothing scheme is constructed according to the properties of the multitensor field estimated using CFARI. The results on synthetic, physical phantom and real brain DW images show that the proposed method is able to better resolve fiber architectures while correctly preserving image edge information, which provides a new tool for investigating the microstructures of biological tissues and for fiber tractography. © 2015 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 25, 285-296, 2015

RHFMs similarity based nonlocal means image denoising in PDTDFB domain

It is a challenging work to design an edge-preserving image denoising scheme. Nonlocal means methods have been widely adopted as simple algorithms for image denoising, but they tend to overblur the image and sharpen the boundary with many texture details lost. In this paper, we propose a new nonlocal means image denoising approach in shiftable complex directional pyramid (PDTDFB) domain, in which the robust radial harmonic Fourier moments (RHFMs) are utilized. The novelty of the proposed approach includes: (1) PDTDFB based multi-scale structural similarity is explored, which has good robustness to noise interference, (2) Geometric invariant radial harmonic Fourier moments (RHFMs) are introduced into similarity measure function, which can improve the denoising ability. Experimental results demonstrate that our image denoising method can obtain better performances in terms of both subjective and objective evaluations than some other state-of-the-art denoising techniques. Especially, it can preserve edges very well while removing noise.

Extended Shearlet HMT Model-Based Image Denoising Using BKF Distribution

Images are often corrupted by noise in the procedures of image acquisition and transmission. It is a challenging work to design an edge-preserving image denoising scheme. Extended discrete Shearlet transform (extended DST) is an effective multi-scale and multi-direction analysis method; it not only can exactly compute the Shearlet coefficients based on a multiresolution analysis, but also can represent images with very few coefficients. In this paper, we propose a new image denoising approach in extended DST domain, which combines hidden Markov tree (HMT) model and Bessel K Form (BKF) distribution. Firstly, the marginal statistics of extended DST coefficients are studied, and their distribution is analytically calculated by modeling extended DST coefficients with BKF probability density function. Then, an extended Shearlet HMT model is established for capturing the intra-scale, inter-scale, and cross-orientation coefficients dependencies. Finally, an image denoising approach based on the extended Shearlet HMT model is presented. Extensive experimental results demonstrate that our extended Shearlet HMT denoising approach can obtain better performances in terms of both subjective and objective evaluations than other state-of-the-art HMT denoising techniques. Especially, the proposed approach can preserve edges very well while removing noise.

A Generalized Detail-Preserving Super-Resolution method

The Super-Resolution (SR) technology, which aims to obtain a high-resolution image by using a set of low-resolution images of the same scene, has become one of the hottest research fields. In this paper, we propose a generalized detail-preserving SR method built on a reasonable observation model and a new image prior model. In order to preserve detail information (i.e., sharp edge and texture information), many SR methods have been established by using the traditional observation models and various image prior models. However, the sensor measurement error, the model error, etc., which are not considered in the existing SR methods, also inevitably make some information get lost from the high-resolution image. In this paper, we use a reasonable observation model that describes the degradation process more fully and exactly for SR reconstruction. Also, we propose an adaptive non-local edge-preserving image prior to model the high-resolution image, which imposes a non-local smoothness constraint on the HR image. Thus, the proposed SR method can better preserve the detail information of an image while avoiding artifacts. The generalized detail-preserving SR method has been tested in artificially generated and real data. The experimental results show that the proposed method can reconstruct higher quality images in both quantitative term and perceptual effect.

Improved L0 Gradient Minimization with L1 Fidelity for Image Smoothing.

Edge-preserving image smoothing is one of the fundamental tasks in the field of computer graphics and computer vision. Recently, L0 gradient minimization (LGM) has been proposed for this purpose. In contrast to the total variation (TV) model which employs the L1 norm of the image gradient, the LGM model adopts the L0 norm and yields much better results for the piecewise constant image. However, as an improvement of the total variation (TV) model, the LGM model also suffers, even more seriously, from the staircasing effect and is not robust to noise. In order to overcome these drawbacks, in this paper, we propose an improvement of the LGM model by prefiltering the image gradient and employing the L1 fidelity. The proposed improved LGM (ILGM) behaves robustly to noise and overcomes the staircasing artifact effectively. Experimental results show that the ILGM is promising as compared with the existing methods.

Efficient multiplicative noise removal method using isotropic second order total variation

To overcome and reduce the undesirable staircase effect commonly met in the total variation (TV) regularization based multiplicative noise removal methods, a novel multiplicative noise removal model based on isotropic second order total variation (ISOTV) is proposed under the maximum a posteriori (MAP) framework. Under the spectral decomposition framework, the ISOTV is first transformed into an equivalent formulation as a novel weighted L1–L2 mixed norm of the second order image derivatives. Then an efficient alternating iterative algorithm is designed to solve the proposed model. Finally, we prove in detail the convergence of the proposed algorithm. A set of experiments on both standard and medical images show that the proposed ISOTV method yields state-of-the-art results both in terms of peak signal to noise ratio (PSNR) and image perception quality. Specifically, the proposed ISOTV method can better reduce the staircase effect and preserve image edges more sharpness with medical applications.

Weighted curvature-preserving PDE image filtering method

The tensor-driven curvature-preserving partial differential equation is an outstanding anisotropic diffusion filtering model. To effectively preserve image edge well, a weighted curvature-preserving PDE based filtering method is proposed, which employ local image directional information to design weight coefficient for different vector fields. Experimental results indicate that new approach shows superior performance on preserving image edge and curvature geometric structure.