Noise-free Image Research Articles

Remote-Sensing Image Denoising With Multi-Sourced Information

In this paper, a new multi-source-information-based method for remote-sensing image denoising is proposed. Different from the conventional denoising algorithm, the proposed method employs the features of similar reference images from the different band, different sensors, or from multi-temporal images when denoising a target image. The noise-free reference images as a prior is introduced into the denoising object function. The prior's information about the reference image is explored in two aspects: dictionary learning and edge-feature prediction. For dictionary learning, we improve the basis training process by incremental singular value decomposition. For edge-feature prediction, we construct the relationship between gradients of the target image and the reference image by linear ridge regression. The new denoising object function employs both the sparsity of the coefficient and the edge similarity between the target image and the reference image. We also present the optimization scheme for the proposed denoising model. Some typical cases based on different feature relations between a target image and a reference image are comprehensively discussed. Reasonably utilizing the similarity between the target image and the reference image, the proposed algorithm smooths out more noise and conserves more detail at the same time. Better performance of the proposed method is confirmed when compared with other state-of-the-art reference-based denoising methods.

End-to-end deep learning framework for digital holographic reconstruction

Digital holography records the entire wavefront of an object, including amplitude and phase. To reconstruct the object numerically, we can backpropagate the hologram with Fresnel–Kirchhoff integral-based algorithms such as the angular spectrum method and the convolution method. Although effective, these techniques require prior knowledge, such as the object distance, the incident angle between the two beams, and the source wavelength. Undesirable zero-order and twin images have to be removed by an additional filtering operation, which is usually manual and consumes more time in off-axis configuration. In addition, for phase imaging, the phase aberration has to be compensated, and subsequently an unwrapping step is needed to recover the true object thickness. The former either requires additional hardware or strong assumptions, whereas the phase unwrapping algorithms are often sensitive to noise and distortion. Furthermore, for a multisectional object, an all-in-focus image and depth map are desired for many applications, but current approaches tend to be computationally demanding. We propose an end-to-end deep learning framework, called a holographic reconstruction network, to tackle these holographic reconstruction problems. Through this data-driven approach, we show that it is possible to reconstruct a noise-free image that does not require any prior knowledge and can handle phase imaging as well as depth map generation.

A new approach for effective retrieval and indexing of medical images

The use of medical imaging has increased manifold in past decade. The medical images have become more and more informative about the patient’s anatomy. Thus, the medical images are now not only used for diagnosis purpose but also stored for R&D purpose in order to have better understanding and deeper insights into the cause and cure of various diseases. For real-time retrieval of medical images from such storage repositories there is a grave need of an effective and efficient biomedical image indexing and retrieval approach. In this quest, this paper presents a new approach for the retrieval of CT and MR images using orthogonal Fourier-Mellin moments (OFMMs). OFMMs have excellent information representation capability that enables them to pack the entire image information in very less number of coefficients. This property makes the proposed approach not only effective but also computationally very efficient and most favorable among all the existing approaches. The proposed approach has been tested and compared with numerous existing, state-of-the-art as well as recently published biomedical indexing and retrieval approaches on four standard databases namely, Emphysema CT, NEMA CT, NEMA MRI and OASIS MRI. Additional experiments have been conducted to analyze the noise robustness ability of the proposed and all the compared approaches. The reported results show a significant increase of about 7% and 20% (average) in the retrieval rate of the proposed approach over all the existing approaches on noise free and noisy images, respectively, of all four test medical databases.

Near-Infrared Cholangiography Can Facilitate a Laparoscopic Choledocholithotomy Even After Upper Abdominal Surgery

Introduction: Surgery for common bile duct (CBD) stones is challenging while unclear hepatic hilum anatomy especial experience of previous laparotomy. A minimally-invasive approach choledocholithotomy can be difficult and converted for the unclear biliary anatomy. Near-infrared (NIR) cholangiography by systemic administration of Indocyanine Green (ICG) can enhance the visualization of the biliary tree anatomy but is limited by the high intensity of background fluorescence signal coming from the liver. NIR fluorescence cholecysto-cholangiography by direct biliary tree administration of the ICG can enhance the biliary tree without background noise signal. We created the NIR cholangiography via different route according to patient situation : systemic circulation or biliary tree injection to see the feasibility of those application Method: Ten patients who suffered from obstructive jaundice due to CBD stone and 5 patients received percutaneous biliary tree drainage as first treatment and 2 patients received endoscopic biliary tree drainage. Those patients received laparoscopic choledocholithotomy as definite treatment after acute infection phase. 5 patients received biliary tree ICG injection via drain tube and 5 patients by systemic injection. Visualization and fluorescence patterns around CBD was recorded Result: In our series, one patient received previous gastrectomy and 4 patients had previous biliary tree surgery. 2 Patients received PTGBD and 3 patient received PTCD for biliary decompression. The near infra-red cholangiography of ten patients can show the clear margin of CBD. No intra or postoperative complications noted. Median time to CBD localization after ICG injection was 5.2 sec for biliary tree injection and 35 minutes for systemic injection. One patient converted to open method due to big stone impaction in distal CBD. No mortality and mobidity were found. Conclusion: Fluorescence cholangiography with ICG injection could provide visualization of CBD even inflammatory or post operative adhesion status, and direct injection method provide more real-time and noise-free image than systemic injection.

A Perceptually Inspired New Blind Image Denoising Method Using $L_{1}$ and Perceptual Loss

Recently, discriminative learning-based denoising methods have received much attention and have been studied to a large extent because of their high denoising performance with significantly shorter inference time compared to model based denoising methods. In this paper, we consider a perceptually motivated blind image denoising problem, which removes various levels of noise from an observed noisy image with a single model and produces visually pleasant images. It is well known that very few blind image denoising methods are available and they have shown the limited quality of restored images because they sacrifice the fine image details during the denoising process due to over-smoothing, resulting in visually unpleasant images. To overcome these problems, we propose a novel loss function that encourages the network to restore noise free images by focusing on the perceived visual quality. Then, the proposed loss function is adopted to an encoder-decoder network with skip connections that produces visually pleasant images with highly preserved fine image details. Further, this method is robust and constantly performs well for several unknown noise levels. Our extensive experimental results show the superiority of the proposed method both quantitatively and qualitatively.

Fast Reconstruction for Monte Carlo Rendering Using Deep Convolutional Networks

Denoising the Monte Carlo (MC) rendering images is different from denoising the natural images since low-sampled MC renderings have a higher noise level and there are inexpensive by-products (e.g., feature buffers) we can leverage. However, the main challenge is designing a model to fast fuse these feature buffers and reconstruct perceptually noise-free images from noisy MC renderings. In recent years, supervised learning methods remove the noise and reconstruct clean images, but most of them cannot handle MC noise well. In this paper, we introduce an end-to-end CNN model to fuse feature buffers and predict a clean image directly. In addition, we devise a new high-dynamic range (HDR) image normalization method to help us to train the model on HDR images in a more efficient and stable way. We setup a series of experiments for selecting the hyperparameter of our deep learning model, network depth, which can promote our network's performance and avoid overfitting. We demonstrate that our model is robust on a wide range of scenes and can generate satisfactory results in a significantly faster way.

Thermographic Laplacian-pyramid filtering to enhance delamination detection in concrete structure

Despite decades of efforts using thermography to detect delamination in concrete decks, challenges still exist in removing environmental noise from thermal images. The performance of conventional temperature-contrast approaches can be significantly limited by environment-induced non-uniform temperature distribution across imaging spaces. Time-series based methodologies were found robust to spatial temperature non-uniformity but requires extended period to collect data. A new empirical image filtering method is introduced in this paper to enhance the delamination detection using blob detection method that originated from computer vision. The proposed method employs a Laplacian of Gaussian filter to achieve multi-scale detection of abnormal thermal patterns by delaminated areas. Results were compared with the state-of-the-art methods and benchmarked with time-series methods in the case of handling non-uniform heat distribution issue. To further evaluate the performance of the method numerical simulations using transient heat transfer models were used to generate the ‘theoretical’ noise-free thermal images for comparison. Significant performance improvement was found compared to the conventional methods in both indoor and outdoor tests. This methodology proved to be capable to detect multi-size delamination using single thermal image. It is robust to non-uniform temperature distribution. The limitations were discussed to refine the applicability of the proposed procedure.

A customized nonlocal restoration schemes with adaptive strength of smoothening for magnetic resonance images

Nonlocal Means (NLM) filter is a nonlinear filter exclusively suitable for images comprising redundant patterns at spatially distinct locations, like Magnetic Resonance (MR) images. However, its performance heavily depends on the selection of one of its arbitrary operational parameters termed as smoothening strength (ξ). Improper selection of ξ may lead to over-blurring or partial denoising. An image adaptive algorithm, to determine the optimum value of ξ, meant for customizing the NLM for MR images is proposed. In the proposed scheme for initializing ξ, its value is increased in proportion to the standard deviation of noise σ^n, according to a linear model ξ=β σ^n. The optimum value of the arbitrary parameter β is identified through an iterative search, using a composite metric termed as Optimum Denoising Index (ODI), which objectively accounts for the noise suppression and edge preservation capability of the filter, as target function. Assuming the noise to be Gaussian distributed with zero mean, its standard deviation is computed using a ‘difference of Laplacian’ kernel. The techniques available in literature for the selection of β are (i) Coupe’s model with β derived out of minimum error sense, (ii) simple linear model ξ=βσ with empirically decided value β = 10 and global optimum of β modified locally in proportion to either (iii) local noise statistics or (iv) edge strength. ODI exhibited by the above techniques and the proposed iterative search are 0.3145 ± 0.0347, 0.2509 ± 0.0149, 0.1210 ± 0.0143, 0.1790 ± 0.0511 and 0.3678 ± 0.0022, respectively. Peak Signal to Noise Ratio (PSNR) between the denoised and the noise-free ground truth images exhibited by Kuwahara Filter, Total Variation (TV) Filter, Anisotropic Diffusion (AD) Filter, Bilateral Filter, SUSAN Filter and the proposed NLM Filter are 21.8739 ± 4.7310, 20.9596 ± 5.0518, 22.1553 ± 5.3369, 22.2142 ± 5.1275, 28.5628 ± 0.02 and 28.9967 ± 0.13, on 30 standard images. The iterative search is superior to methods available in literature with regard to the sharpness of the true morphological edges and smoothness of the homogenous regions in the denoised image. The proposed scheme of NLM is found to be superior to Kuwahara, TV, AD, Bilateral and SUSAN Filters.

Simple image artifact removal technique for more accurate iris diagnosis

Iris diagnosis based on the color and texture information is one of a novel approach which can represent the current state of a certain organ inside body or the health condition of a person. In analysis of the iris images, there are critical image artifacts which can prevent of use interpretation of the iris textures on images. Here, we developed the iris diagnosis system based on a hand-held typed imaging probe which consists of a single camera sensor module with 8M pixels, two pairs of 400~700 nm LED, and a guide beam. Two original images with different light noise pattern were successively acquired in turns, and the light noise-free image was finally reconstructed and demonstrated by the proposed artifact removal approach.

Simultaneous image fusion and denoising by using fractional-order gradient information

Image fusion and denoising are significant in image processing because of the availability of multi-sensor and the presence of the noise. The first-order and second-order gradient information have been effectively applied to deal with fusing the noise-free source images. In this paper, we utilize the fractional-order derivatives to represent image features, and propose two new convex variational models for fusing noisy source images. Furthermore, we apply an alternating direction method of multiplier (ADMM) to solve the minimization problems in the proposed models. Numerical experiments show that the proposed methods outperform the conventional total variation methods for simultaneously fusing and denoising.

Feature-oriented singular value shrinkage for optical coherence tomography image

Optical coherence tomography (OCT) is a non-invasive optical imaging modality that has been widely used in the field of medical diagnosis. However, OCT images are often degraded by speckle noise. To address this problem, this paper proposes a two-stage feature-oriented singular value shrinkage algorithm in a low-rank approximation framework, for speckle noise reduction and contrast enhancement of intra-retinal layers of OCT images. First, a weighted absolute distance is employed to find nonlocal similar patches that exhibit high correlation to a given reference one. Next, the singular values of the group matrix formed by similar patches are shrunk by mixed thresholding so that they are closer to the singular values of that of latent noise-free image. Finally, an iterative regularization technique is adopted to improve the denoising performance of the proposed method, where the backward projection parameter in each pixel is adaptively determined by its corresponding gradient variation. By using these strategies, the proposed method not only effectively removes the speckle noise of OCT images, but also preserves fine structural information of objects. Experimental results show that the proposed algorithm is competitive with some state-of-the-art speckle removal techniques in terms of both objective metrics and subjective visual inspection.

Suspicious Lesion Segmentation on Brain, Mammograms and Breast MR Images Using New Optimized Spatial Feature Based Super-Pixel Fuzzy C-Means Clustering.

Suspicious lesion or organ segmentation is a challenging task to be solved in most of the medical image analyses, medical diagnoses and computer diagnosis systems. Nevertheless, various image segmentation methods were proposed in the previous studies with varying success levels. But, the image segmentation problems such as lack of versatility, low robustness, high complexity and low accuracy in up-to-date image segmentation practices still remain unsolved. Fuzzy c-means clustering (FCM) methods are very well suited for segmenting the regions. The noise-free images are effectively segmented using the traditional FCM method. However, the segmentation result generated is highly sensitive to noise due to the negligence of spatial information. To solve this issue, super-pixel-based FCM (SPOFCM) is implemented in this paper, in which the influence of spatially neighbouring and similar super-pixels is incorporated. Also, a crow search algorithm is adopted for optimizing the influential degree; thereby, the segmentation performance is improved. In clinical applications, the SPOFCM feasibility is verified using the multi-spectral MRIs, mammograms and actual single spectrum on performing tumour segmentation tests for SPOFCM. Ultimately, the competitive, renowned segmentation techniques such as k-means, entropy thresholding (ET), FCM, FCM with spatial constraints (FCM_S) and kernel FCM (KFCM) are used to compare the results of proposed SPOFCM. Experimental results on multi-spectral MRIs and actual single-spectrum mammograms indicate that the proposed algorithm can provide a better performance for suspicious lesion or organ segmentation in computer-assisted clinical applications.

A new compressive sensing based image denoising method using block-matching and sparse representations over learned dictionaries

Suppressing noise and preserving detail information such as edges and textures are two key challenges in image denoising. In this paper, a new method for eliminating noise from images is presented which is based on not only compressive sensing but also sparse and redundant representations over trained dictionaries. The objective function of the proposed technique consists of two terms. The first term processes the noisy image by the hard thresholding operator in the bandelet domain to provide the noise-free image as well as guaranteeing the similarity between the denoised image and the noisy image, while the second term ensures that the image admits a sparse decomposition in a dictionary. In addition, the proposed method takes advantage of the block-matching technique for representing the dictionary elements such that the noisy image is firstly grouped by the block-matching technique, and then an identical sparse vector is used for all patches in a group. Simulations using images contaminated by additive white Gaussian noise demonstrate that the performance of the proposed method considerably surpasses that of state-of-the-art methods, both visually and in terms of quantitative criteria, namely peak signal to noise ratio and structural similarity.

A Boosting SAR Image Despeckling Method Based on Non-Local Weighted Group Low-Rank Representation.

In this paper, we propose a boosting synthetic aperture radar (SAR) image despeckling method based on non-local weighted group low-rank representation (WGLRR). The spatial structure information of SAR images leads to the similarity of the patches. Furthermore, the data matrix grouped by the similar patches within the noise-free SAR image is often low-rank. Based on this, we use low-rank representation (LRR) to recover the noise-free group data matrix. To maintain the fidelity of the recovered image, we integrate the corrupted probability of each pixel into the group LRR model as a weight to constrain the fidelity of recovered noise-free patches. Each single patch might belong to several groups, so different estimations of each patch are aggregated with a weighted averaging procedure. The residual image contains signal leftovers due to the imperfect denoising, so we strengthen the signal by leveraging on the availability of the denoised image to suppress noise further. Experimental results on simulated and actual SAR images show the superior performance of the proposed method in terms of objective indicators and of perceived image quality.

Evolutionary multiobjective multiple description wavelet based image coding in the presence of mixed noise in images

In this paper, a novel method for generation of multiple description (MD) wavelet based image coding is proposed by using Multi-Objective Evolutionary Algorithms (MOEAs). Complexity of the multimedia transmission problem has been increased for MD coders if an input image is affected by any type of noise. In this case, it is necessary to solve two different problems which are designing the optimal side quantizers and estimating optimal parameters of the denoising filter. Existing MD coding (MDC) generation methods are capable of solving only one problem which is to design side quantizers from the given noise-free image but they can fail reducing any type of noise on the descriptions if they applied to the given noisy image and this will cause bad quality of multimedia transmission in networks. Proposed method is used to overcome these difficulties to provide effective multimedia transmission in lossy networks. To achieve it, Dual Tree-Complex Wavelet Transform (DT-CWT) is first applied to the noisy image to obtain the subbands or set of coefficients which are used as a search space in the optimization problem. After that, two different objective functions are simultaneously employed in the MOEA to find pareto optimal solutions with the minimum costs by evolving the initial individuals through generations. Thus, optimal quantizers are created for MDCs generation and obtained optimum parameters are used in the image filter to remove the mixed Gaussian impulse noise on the descriptions effectively. The results demonstrate that proposed method is robust to the mixed Gaussian impulse noise, and offers a significant improvement of optimal side quantizers for balanced MDCs generation at different bitrates.

Morphological Contour Based Blood Vessel Segmentation in Retinal Images Using Otsu Thresholding

Retinal image analysis plays an important part in identifying various eye related diseases such as diabetic retinopathy (DR), glaucoma and many others. Accurate segmentation of blood vessels plays an important part in identifying the retinal diseases at an early stage. In this article, an unsupervised approach based on contour detection has been proposed for effective segmentation of retinal blood vessels. The proposed morphological contour-based blood vessel segmentation (MCBVS) method performs preprocessing using contrast limited adaptive histogram equalization followed by alternate sequential filtering to generate a noise-free image. The resultant image undergoes Otsu thresholding for candidate extraction followed by contour detection to properly segment the blood vessels. The MCBVS method has been tested on the DRIVE dataset and the experimental result shows that the proposed method achieved a sensitivity, specificity and accuracy of 58.79%, 90.77% and 86.7%, respectively. The MCBVS method performs better than the existing methods Sobel, Prewitt and Modified U-Net in terms of accuracy.

Mixed Gaussian-impulse noise reduction from images using convolutional neural network

The removal of mixed-noise is an ill-posed problem due to high level of non-linearity in the distribution of noise. Most commonly encountered mixed-noise is the combination of additive white Gaussian noise (AWGN) and impulse noise (IN) that have contrasting characteristics. A number of methods from the cascade of IN and AWGN reduction to the state-of-the-art sparse representation have been reported to reduce this common form of mixed-noise. In this paper, a new learning-based algorithm using the convolutional neural network (CNN) model is proposed to reduce the mixed Gaussian-impulse noise from images. The proposed CNN model adopts computationally efficient transfer learning approach to obtain an end-to-end map from noisy image to noise-free image. The model has a small structure yet it is capable of providing performance superior to that of the well established methods. Experimental results on different settings of mixed-noise show that the proposed CNN-based denoising method performs significantly better than the sparse representation and patch-based methods do both in terms of accuracy and robustness. Moreover, due to the lightweight structure, the denoising operation of the proposed CNN-based method is computationally faster than that of the previously reported methods.

Blind deblurring using coupled convolutional sparse coding regularisation for noisy‐blurry images

This Letter proposes a novel method to deblur a blurry image corrupted by noise. The authors estimate a noise-free version of the input blurred image and a corresponding noise-free version of the latent image without damaging the blur information, as well as the latent image and blur kernel in an alternating fashion. To this end, they first propose coupled convolutional sparse coding, which incorporates the coupled dictionary concept into convolutional sparse coding. Then they model the noise-free blurred image to share the sparse coefficients with the noise-free latent image using the coupled dictionaries. By utilising these noise-free images as priors in alternating latent image estimation and blur kernel estimation steps, they can estimate a high-quality latent image and blur kernel in the presence of noise. Experimental results demonstrate that the proposed method outperforms previous methods in handling noisy blurred images.

A New ROI-Based performance evaluation method for image denoising using the Squared Eigenfunctions of the Schrödinger Operator.

In this paper a new Region Of Interest (ROI) characterization for image denoising performance evaluation is proposed. This technique consists of balancing the contrast between the dark and bright ROIs, in Magnetic Resonance (MR) images, to track the noise removal. It achieves an optimal compromise between removal of noise and preservation of image details. The ROI technique has been tested using synthetic MRI images from the BrainWeb database. Moreover, it has been applied to a recently developed denoising method called Semi-Classical Signal Analysis (SCSA). The SCSA decomposes the image into the squared eigenfunctions of the Schrödinger operator where a soft threshold $h$ is used to remove the noise. The results obtained using real MRI data suggest that this method is suitable for real medical image processing evaluation where the noise-free image is not available.

Automated spectral domain approach of quasi‐periodic denoising in natural images using notch filtration with exact noise profile

The domain of noise fading from digital images, by virtue of its enormous appellation amongst the researchers, stands out uniquely in the recent research field of image processing over the last few decades. Periodic noises are unintended spurious signals which often agitate an image during acquisition/transmission, thereby resulting in repetitive patterns having spatial dependency and extensively demeaning visual excellence of the image. However, high amplitude noisy spectral components are clearly noticeable from the remaining uncorrupted ones in the corresponding Fourier transformed corrupted image spectrum. Hence, it is easier to distinguish and minimise those noisy components using an appropriate thresholding and filtration technique. Therefore, to start with, a simple yet elegant model of the noise-free natural image has been developed from the corrupted one followed by a proper thresholding method to get the noisy bitmap. Finally, an elegant adaptive sinc restoration filter with the concept of extracting the exact shape of a noise spectrum profile has been applied in the filtration phase. The performance of the proposed algorithm has been assessed both visually and statistically with other state-of-the-art algorithms in the literature in terms of various performance measurement attributes, providing evidence of achieving more effective restoration with considerable lower computational time.