Denoising Of Digital Images Research Articles

An intelligent lithology recognition system for continental shale by using digital coring images and convolutional neural networks

Lithology identification is a critical component of reservoir evaluation and hydrocarbon development. However, the continental shale rocks, characterized by complex lithologies and developed laminations, present new challenges to traditional lithology identification techniques. In this study, we develop a centimeter-level intelligent identification system for continental shale lithology, leveraging digital image preprocessing and deep learning methods. Specifically, we construct a dataset from 400 m of drilling cores, encompassing 24 types of lithology and 152 lithological transitions. The downsampling and oversampling methods are employed to balance the dataset. The digital image denoising and restoration techniques are then utilized to enhance rock image features. Different neural network models are examined to obtain the optimum predictive results. The results indicate that data balancing improves the TOP1 and TOP5 accuracies by 10% and 5%, respectively. Gaussian low-pass filter denoising, in comparison to other methods, increases the TOP1 and TOP5 accuracies by 2%, establishing it as the superior denoising technique for lithology recognition. Considering both computational efficiency and prediction accuracy, EfficientNet emerges as the most effective model for continental shale image recognition. The optimized model attains up to 60% TOP1 and 95% TOP5 across 22 categories in drilling core samples. This study delves into the intelligent identification of macroscopic drilling core images of continental shale, and the findings can offer technical support for the interpretation and evaluation of continental shale reservoirs.

Security Protection Mechanism in Cloud Computing Authorization Model Using Machine Learning Techniques

Because of the vast number of applications and the ambiguity in application methods, handwritten character recognition has garnered widespread recognition and increased prominence in the community of pattern recognition researchers ever since it was first developed. This is due to the fact that application methods can be quite ambiguous. Computer in the cloud, on the other hand, allows for suitable network access on demand to a shared pool of customizable computing resources and digital devices. According to those knowledgeable in the subject, the standard filtering techniques are not enough when it comes to the process of denoising images. In many different approaches to machine learning, information is lost not just during the filtering process itself but also at other points during the process. When a convolutional neural network is going through its pooling operation, the internal data representation either becomes misaligned or entirely vanishes (CNN). The reconstruction of low-intensity digital photographs, which takes place during repetitive filtering, breaks away the artefacts that remain after each filtering function, which results in an image that is more uniform. The multilayer wavelet transform, or MLWT, is a method for processing features that comprises of many filter bands and is used in cloud computing authorization that is protected securely. In this scenario, a significant quantity of information gets obliterated from digital photographs during the process of feature extraction and processing. These issues are investigated by the deep learning algorithms that make use of autoencoder, and the methods also handle the novel windowing blocks that are being introduced to the layers. In this section, the magnitude and phase information is considered in order to construct a deep learning framework that will provide good denoising of digital images. The proposed architecture is equipped with the capability of accurately identifying, in real time, the noise level and type that was employed in the training of the network. The method that we have proposed, which is centred on the distribution of noise, may be used to determine the kind of noise. In order to categorise the various types of noise, we investigated nine distinct noise distributions. Dilated convolutional filtering will be used as the method of choice in order to ascertain the specific nature of the noise that can be found in the digital images. An autoencoder-based deep learning algorithm is able to accomplish numerous experimental results in digital image denoising operations that are superior to those achieved by a typical deep learning algorithm even when the intensity of the scenario is low. The performance of the method that is produced by combining the autoencoder and the dilated convolutional filtering techniques is enhanced over the performance of the technology that is currently in use. Using the method that we have outlined in this study, it is possible to recreate the low-intensity images in their entirety. We were able to show that our proposed method beat other existing algorithms for low-density digital photos by comparing several metrics, such as the peak signal-to-noise ratio (PSNR), the structural similarity index, and others (SSIM).

DIGITAL IMAGE RESTORATION USING LINEAR PDE-BASED FILTERING MODELS

The linear partial differential equation (PDE) - based models for Gaussian noise removal are discussed in this paper. The digital image denoising and restoration solutions based on linear diffusion equations are surveyed first. Then, our own contributions in this image processing domain, representing some effective linear PDE-based filtering models based on hyperbolic and stochastic differential equations, are presented. The results of our denoising experiments are also provided in this article.

Optimal reduction of noise in image processing using collaborative inpainting filtering with Pillar K-Mean clustering

ABSTRACTDigital image processing is a mechanism for analysing and modifying the image in order to improve the quality and also to manage the unwanted involvement of noises. In image processing, noise is characterized as an unwanted disturbance which occurs while capturing the actual image thus affecting the quality of the image. Hence, noise formation is considered as a perilous issue and the reduction of noise is considered as an awkward process. Nowadays, almost in all fields of science and technology, digital image processing is increasing rapidly, so there arises the need for de-noising to cure the noised image. The main objective of this paper is to overcome the issue of noise and also to increase the quality and pixel value of the image. An advanced methodology known as collaborative filtering and Pillar K-Mean clustering is discussed in this paper to overcome the abovementioned problem. Initially, distinct pure images are taken as the dataset and three types of noises are added to the corresponding image to make it as a noised one. Hence, the unspecified noise is resolved on the basis of a hybrid combination of algorithms of collaborative filtering with the image inpainting method. Sequentially, the low-density noises, such as random noise and poison noise, are recovered by the implementation of collaborative filtering, and the high-density salt and pepper noise are recovered by the image inpainting method. Based on the GLCM (Grey Level Co-occurrence Matrix) feature, the normal image and the noised image are used for the clustering process. Then the de-noised image is evaluated to find the efficiency on the basis of few parameters such as SNR (Signal to Noise Ratio), MSE (Mean Square Error), PSNR (Peak Signal to Noise Ratio) and SSI (Structural Similarity Index). Accordingly, the evaluated images are further withstood for clustering to differentiate the noises by applying the proposed clustering methodology. Then the evaluated images are verified on the basis of a few parameters such as Silhouette Width, Davies–Bouldin Index and Dunn Index. The proposed methodology is run on the platform of Mat Lab. Finally, the proposed methodology is considered as an efficient method for settling the issue in digital image de-noising.

Interview

Janarthanam Subramaniam Janarthanam Subramaniam from the National Institute of Electronics and Information Technology, India, talks to us about the work behind his paper ‘Fast median-finding word comparator array’, page 1402. I consider myself blessed to have the opportunity to work with Prof. David Ebenezer, who inspired me in the ethical values of research and provided guidance on publishing work. I began my research developing tunnelling current models for MOSFET and Carbon Nanotube Field Effect Transistors. Under the guidance of Prof. Ebenezer, I went on to develop nonlinear digital filters. We are interested in digital image processing, de-noising in particular. Impulse noise is not amenable to standard mathematics, physics, and linear system theory. Hence, the interest in median filtering, which is deceptively simple but does what linear system theory fails to do. There are a good number of publications in median filtering of digital images, however, hardware implementation has not taken off effectively. This is the motivation behind our Letter. We aspire to make median filter the standard de-noising block in the Image Signal Processing (ISP) environment. Pipelining and parallelism is a part of the Harvard Architecture, which is a standard platform for real-time digital processing. However, this architecture by itself is not sufficient for issues such as latency and other hardware resources. There is a need for continued systematic empirical search, which will provide enhanced support to pipelining and parallelism. The issue is non-availability of theoretical frame work. In our Letter we present a pipelined architecture of fast median-finding word comparator array. Effective parallel schemes for hardware implementation are the challenging part of hardware architecture development in signal processing system design. Most of the non-sorting methods are iterative but have smaller latency. Sorting methods are non-iterative but lead to higher latency. Under sorting methods, our proposed method addresses reduction in latency. We have reported an eight stage systolic array for an efficient median filtering technique which uses a 3×3 mask. The efficiency implies the reduction in the number of comparator stages, which results in reduced latency and fewer number of signal paths in the word comparator array. The method uses selective comparators for finding the median by avoiding comparators required to obtain fully sorted lists. The reduction in latency is the novelty of the proposed technique. This reduction is significant in the context of ISP, for example, in digital photography, cinematography, forensic science and electronic surveillance using technology such as drones. The reported method can be extended to signed integers of any data width. Non-sorting based methods in the literature involve arithmetic operations. Though new methods of median finding are being developed, extensions with parallelism and pipelining find very limited consideration. This makes it difficult to categorise the methods and related applications in hardware implementation. The most difficult challenge was the lack of theoretical support for predictable methods for efficiently sequencing comparator operations in time. This required laborious and time consuming empirical searching. De-noising in real-time applications such as electronic surveillance, digital videography and critical care biomedical electronics are the target areas for applications of the proposed method. I am working on hardware implementation of algorithms for de-noising of digital images and developing parallel and pipelined architectures for different state-of-the-art median finding methods. Since the National Institute of Electronics & IT is an electronic product design institute, it provides opportunities for us to develop and prototype algorithms for computer vision environments. Also, we are working on medical image processing for automated diagnostic system design and quality improvements. Newly developed nonlinear digital filters are being validated with state-of-the-art de-noising techniques on system level. I expect to see the development of reconfigurable nonlinear digital filters to support decision based filtering requirements. The development of hardware architecture for decision-based median filters and the design of circuits in VLSI in the context of ISP are the long-term goals. Support of reconfigurable hardware, low power VLSI and high processing capabilities in the near future will provide added advantages for achieving more significant results in this rare area of research.

Performance Study of Novel Median Filter for Image De-Noising

Most common method of clearing image noise is through median filtration. Median smoothers discard outliers (or impulses) quite effectively at lower noise densities but, fail to provide adequate smoothing for heavily noised images for the simple reason that, in the conventional median filter approaches, each and every pixel of an image is filtered without concern about healthy pixels. To suppress this deficiency, in this paper, an improved median filter by the name Novel Median Filter (NMF) is proposed to enhance the effectiveness of the median filter at higher noise situations to provide a detail preserving de-noising of digital images. Proposed NMF algorithm is a 3-stage filter; first two stages are dedicated to selective switching median filtration in which noisy pixels are first detected and then the filtering operation is applied to restore their values with a valid (or a noise free) median computed for each 3×3 window position. Third level processing enables to remove any distortions present in the de-noised image. The proposed method, Novel Median Filter (NMF) is evaluated for its effectiveness in detail preserving de-noising capability on a large number of digital grey and color images affected with impulsive and non-impulsive noise in terms of objective and subjective quality measures. To better appraise the efficacy of the proposed method, Novel Median Filter (NMF), MATLAB simulation results in terms of peak signal to noise ratio (PSNR), mean square error (MSE) and the de-noised images are obtained and are compared with those obtained for a standard median filter (SMF) and its few variants and meaningful conclusions are derived. Detail preservation capability of the proposed filter has been demonstrated through experimentation on medical samples. General Terms Image Processing, Nonlinear Filter, Median Filtration.

EBMBDT: Effective Block Matching Based Denoising Technique using dual tree complex wavelet transform

In processing and investigation of digital image denoising of images is hence very important. In this paper, we propose a Hybrid de-noising technique by using Dual Tree Complex Wavelet Transform (DTCWT) and Block Matching Algorithm (BMA). DTCWT and BMA is a method to identify the noisy pixel information and remove the noise in the image. The noisy image is given as input at first. Then, bring together the comparable image blocks into the load. Afterwards Complex Wavelet Transform (CWT) is applied to each block in the group. The analytic filters are made use of by CWT, i.e. their real and imaginary parts from the Hilbert Transform (HT) pair, defending magnitude-phase representation, shift invariance, and no aliasing. After that, adaptive thresholding is applied to enhance the image in which the denoising result is visually far superior. The proposed method has been compared with our previous de-noising technique with Gaussian and salt-pepper noise. From the results, we can conclude that the proposed de-noising technique have shown better values in the performance analysis.

Digital Image Denoising by Adaptive Median Filtering and Wavelet Transform

Digital image processing is an important area of preservation and representation of cultural and historical objects. Part of it is the task of cleaning images corrupted by noise. Many different methods and approaches have been successfully implemented, but still filtering and denoising techniques are developing and improving intensively. Herein, we report numeric simulations of 2D image filtering. Effects of impulse noise are very efficiently reduced by adaptive median filtering. The situation of additive noise, which itself depends on the intensity of the original image, is more complicated. However, we demonstrate that in this case the use of wavelet transforms in the denoising procedure is very productive and promising.

Denoising of digital images through PSO based pixel classification

Abstract This paper proposes a de-noising method where the detection and filtering is based on unsupervised classification of pixels. The noisy image is grouped into subsets of pixels with respect to their intensity values and spatial distances. Using a novel fitness function the image pixels are classified using the Particle Swarm Optimization (PSO) technique. The distance function measured similarity/dissimilarity among pixels using not only the intensity values, but also the positions of the pixels. The detection technique enforced PSO based clustering, which is very simple and robust. The filtering operator restored only the noisy pixels keeping noise free pixels intact. Four types of noise models are used to train the digital images and these noisy images are restored using the proposed algorithm. Results demonstrated the effectiveness of the proposed technique. Various benchmark images are used to produce restoration results in terms of PSNR (dB) along with other parametric values. Some visual effects are also presented which conform better restoration of digital images through the proposed technique.

Weighted and well-balanced anisotropic diffusion scheme for image denoising and restoration

Anisotropic diffusion is a key concept in digital image denoising and restoration. To improve the anisotropic diffusion based schemes and to avoid the well-known drawbacks such as edge blurring and ‘staircasing’ artifacts, in this paper, we consider a class of weighted anisotropic diffusion partial differential equations (PDEs). By considering an adaptive parameter within the usual divergence process, we retain the powerful denoising capability of anisotropic diffusion PDE without any oscillating artifacts. A well-balanced flow version of the proposed scheme is considered which adds an adaptive fidelity term to the usual diffusion term. The scheme is general, in the sense that, different diffusion coefficient functions can be utilized according to the need and imaging modality. To illustrate the advantage of the proposed methodology, we provide some examples, which are applied in restoring noisy synthetic and real digital images. A comparison study with other anisotropic diffusion based schemes highlight the superiority of the proposed scheme.

On a family of high-order iterative methods under gamma conditions with applications in denoising

We study a class of at least third order iterative methods for nonlinear equations on Banach spaces. A characterization of the convergence under Gamma-type conditions is presented. Though, in general, these methods are not very extended due to their computational costs, we can find examples in which they are competitive and even cheaper than other simpler methods. Indeed, we propose a new nonlinear mathematical model for the denoising of digital images, where the best method in the family has fourth order of convergence. Moreover, our family includes two-step Newton type methods with good numerical behavior in general. We center our analysis in both, analytic and computational, aspects.

Box Relaxation Schemes in Staggered Discretizations for the Dual Formulation of Total Variation Minimization

In this paper, we propose some new box relaxation numerical schemes on staggered grids to solve the stationary system of partial differential equations arising from the dual minimization problem associated with the total variation operator. We present in detail the numerical schemes for the scalar case and its generalization to multichannel (vectorial) images. Then, we discuss their implementation in digital image denoising. The results outperform the resolution of the dual equation based on the gradient descent approach and pave the way for more advanced numerical strategies.

Functional Link Artificial Neural Network for Denoising of Image

Digital image denoising is crucial part of image preprocessing. The application of denoising procesn satellite image data and also in television broadcasting. Image data sets collected by image sensors are generally contaminated by noise. Furthermore, noise can be introduced by transmission errors and compression. Thus, denoising is often a necessary and the first step to be taken before the images data is analyzed. In this paper a Modified Functional Link Artificial Neural Network (M-FLANN) is proposed which is simpler than a Multilayer Perceptron (MLP). It have been implemented for image restoration in this paper. Its computational complexity and speed and generalization ability to cancel Gaussian noise is compared with that of MLP. In the single layer functional link ANN (FLANN) the need of hidden layer is eliminated. The novelty of the FLANN structure is that it requires much less computation than that of MLP. In the presence of additive white Gaussian noise, salt and pepper noise, Random variable impulse noise and mixed noise in the image the performance of the proposed network is compared with that of MLP in this paper. The Performance of the of algorithm is evaluated for six different situations i.e. for single layer neural network, MLP and four different types of expansion in FLANN and comparison in terms of computational complexity also carried out. Index Terms— MLP, FLANN, Salt and Pepper noise.

A well-posed multiscale regularization scheme for digital image denoising

A well-posed multiscale regularization scheme for digital image denoisingWe propose an edge adaptive digital image denoising and restoration scheme based on space dependent regularization. Traditional gradient based schemes use an edge map computed from gradients alone to drive the regularization. This may lead to the oversmoothing of the input image, and noise along edges can be amplified. To avoid these drawbacks, we make use of a multiscale descriptor given by a contextual edge detector obtained from local variances. Using a smooth transition from the computed edges, the proposed scheme removes noise in flat regions and preserves edges without oscillations. By incorporating a space dependent adaptive regularization parameter, image smoothing is driven along probable edges and not across them. The well-posedness of the corresponding minimization problem is proved in the space of functions of bounded variation. The corresponding gradient descent scheme is implemented and further numerical results illustrate the advantages of using the adaptive parameter in the regularization scheme. Compared with similar edge preserving regularization schemes, the proposed adaptive weight based scheme provides a better multiscale edge map, which in turn produces better restoration.

A Digital Diffusion-Reaction Type Filter for Nonlinear Denoising

We propose a new digital diffusion-reaction filter for denoising of digital images and data living on graphs. The construction of this filter is motivated by the regularization and diffusion approach to edge detection by Nordstrom. Properties and convergence of the iteration scheme are studied and numerical results are given.

THE DYADIC LIFTING SCHEMES AND THE DENOISING OF DIGITAL IMAGES

The dyadic lifting schemes, which generalize Sweldens lifting schemes, have been proposed for custom-design of dyadic and bi-orthogonal wavelets and their duals. Starting with dyadic wavelets and exploiting the control provided in the form of free parameters, one can custom-design dyadic as well as bi-orthogonal wavelets adapted to a particular application. To validate the usefulness of the schemes, two construction methods have been proposed for designing dyadic wavelet filters with higher number of vanishing moments; using these design techniques, spline dyadic wavelet filters have been custom-designed for denoising of digital images, which exhibit enhanced denoising effects.

Feature-based wavelet shrinkage algorithm for image denoising

A selective wavelet shrinkage algorithm for digital image denoising is presented. The performance of this method is an improvement upon other methods proposed in the literature and is algorithmically simple for large computational savings. The improved performance and computational speed of the proposed wavelet shrinkage algorithm is presented and experimentally compared with established methods. The denoising method incorporated in the proposed algorithm involves a two-threshold validation process for real-time selection of wavelet coefficients. The two-threshold criteria selects wavelet coefficients based on their absolute value, spatial regularity, and regularity across multiresolution scales. The proposed algorithm takes image features into consideration in the selection process. Statistically, most images have regular features resulting in connected subband coefficients. Therefore, the resulting subbands of wavelet transformed images in large part do not contain isolated coefficients. In the proposed algorithm, coefficients are selected due to their magnitude, and only a subset of those selected coefficients which exhibit a spatially regular behavior remain for image reconstruction. Therefore, two thresholds are used in the coefficient selection process. The first threshold is used to distinguish coefficients of large magnitude and the second is used to distinguish coefficients of spatial regularity. The performance of the proposed wavelet denoising technique is an improvement upon several other established wavelet denoising techniques, as well as being computationally efficient to facilitate real-time image-processing applications.