Search

Recently, image segmentation based on graph cut methods has shown remarkable performance on a set of image data. Although the kernel graph cut method provides good performance, its performance is highly dependent on the data mapping to the transformation space and image features. The entropy-based kernel graph cut method is suitable for segmentation of textured images. Nonetheless, its segmentation quality remains significantly contingent on the accuracy and richness of feature space representation and kernel centers. This paper introduces an entropy-based kernel graph cut method, which leverages the discriminative feature space extracted from SqueezeNet, a deep neural network. The fusion of SqueezeNet’s features enriches the segmentation process by capturing high-level semantic information. Moreover, the extraction of kernel centers is refined through a weighted k-means approach, contributing further to the segmentation’s precision and effectiveness. The proposed method, while exploiting the benefits of suitable computational load of graph cut methods, will be a suitable alternative for segmenting textured images. Laboratory results have been taken on a set of well-known datasets that include textured shapes in order to evaluate the efficiency of the algorithm compared to other well-known methods in the field of kernel graph cut. [ABSTRACT FROM AUTHOR]

Image restoration has been a longstanding challenge in image processing, involving the enhancement of image content to extract valuable information. Nonlinear models have demonstrated their efficacy in eliminating additive noise, which motivates our goal of investigating a novel nonlocal nonlinear reaction-diffusion model for noise removal. The proposed model integrates a fractional diffusion equation with a nonlinear nonlocal p -Laplace operator, with the fidelity term employing a weak norm to better capture oscillatory patterns and intricate details in textured images. Using the Schauder fixed point theorem, the well-posedness of the proposed model solution is established. The experimental results confirm the effectiveness and efficiency of the proposed model, providing validation for its practical utility. [ABSTRACT FROM AUTHOR]

A new nonlocal total variation reaction–diffusion system is proposed, analysed and implemented for image denoising. In the way, it is obtained by analysing the case where p → 1 in the nonlocal p-Laplacian reaction–diffusion system proposed in [F. Karami, D. Meskine, and K. Sadik, A new nonlocal model for the restoration of textured images, J. Appl. Anal. Comput. 9(6) (2019), pp. 2070–2095.]. We established the existence and uniqueness of the proposed model. Numerical and comparatives experiments are presented that illustrate the main features of the new model and that demonstrate its effectiveness as a denoising tool. [ABSTRACT FROM AUTHOR]

This paper addresses the analysis of textured images using the symmetric positive semi-definite matrix. In particular, a field of symmetric positive semi-definite matrices is used to estimate the structural information represented by the local orientation and the degree of anisotropy in structured and sinusoid-like textured images. In order to ensure faithful local structure estimation, an adaptive algorithm for the regularization of the extent of gradient fields smoothing is proposed. Results obtained on different texture samples show the strength of the proposed method in accurately representing the local variation of orientations in the underlying textured images, which paves the way towards an accurate analysis of the texture structures.

In this paper, we focus on the mathematical and numerical study of a variable exponent nonlocal p ( X ) -Laplacian equation for image denoising. Based on the Semigroup Theory, we prove the existence and uniqueness of solution for the proposed model. To illustrate the efficiency and effectiveness of our model, we provide the denoising experimental results as well we compare it with some existing models in the literature. [ABSTRACT FROM AUTHOR]

This letter addresses an estimation problem based on blurred and noisy observations of textured images. The goal is jointly estimating the 1) image model parameters, 2) parametric point spread function (semi-blind deconvolution) and 3) signal and noise levels. It is an intricate problem due to the data model non-linearity w.r.t. these parameters. We resort to an optimal estimation strategy based on Mean Square Error, yielding the best (non-linear) estimate, namely the Posterior Mean. It is numerically computed using a Monte Carlo Markov Chain algorithm: Gibbs loop including a Random Walk Metropolis-Hastings sampler. The novelty is double: i) addressing this fully parametric threefold problem never tackled before through an optimal strategy and ii) providing a theoretical Fisher information-based analysis to anticipate estimation accuracy and compare with numerical results. [ABSTRACT FROM AUTHOR]

Highlights: [•] We present an objective performance analysis of statistical edge detection. [•] A controlled dataset of images and tests is given and fully analysed. [•] We show how statistical tests can outperform traditional edge detection methods. [•] We present recommendations for the tests, giving guidelines for their use. [•] Non-parametric tests perform best overall, giving improvements over all tests. [Copyright &y& Elsevier]

Abstract: In this paper we propose a new method for combining simple classifiers through the analogue Hopfield Neural Network (HNN) optimization paradigm for classifying natural textures in images. The base classifiers are the Fuzzy clustering (FC) and the parametric Bayesian estimator (BP). An initial unsupervised training phase determines the number of clusters and estimates the parameters for both FC and BP. Then a decision phase is carried out, where we build as many Hopfield Neural Networks as the available number of clusters. The number of nodes at each network is the number of pixels in the image which is to be classified. Each node at each network is initially loaded with a state value, which is the membership degree (provided by FC) that the node (pixel) belongs to the cluster associated to the network. Each state is later iteratively updated during the HNN optimization process taking into account the previous states and two types of external influences exerted by other nodes in its neighborhood. The external influences are mapped as consistencies. One is embedded in an energy term which considers the states of the node to be updated and the states of its neighbors. The other is mapped as the inter-connection weights between the nodes. From BP, we obtain the probabilities that the nodes (pixels) belong to a cluster (network). We define these weights as a relation between states and probabilities between the nodes in the neighborhood of the node which is being updated. This is the classifier combination, making the main finding of this paper. The proposed combined strategy based on the HNN outperforms the simple classifiers and also classical combination strategies. [Copyright &y& Elsevier]

Abstract: This paper presents a new unsupervised hybrid classifier that combines several base classifiers through a fuzzy multicriteria decision making (MCDM) approach. The base classifiers are: fuzzy clustering, parametric and non-parametric Bayesian approaches, self-organizing feature maps and two versions of learning vector quantization. During the learning phase different partitions are established until a valid partition is found. The partitioning and validation are two automatic processes based on validation measurements. These measures allow computing the competences of each base classifier which are mapped as the weights to be used during the decision process through the MCDM. The design of the unsupervised classifier from supervised base classifiers and the automatic computation of the competences make the main contributions of this paper. Although the method is designed for six classifiers it can be extended for a greater number of classifiers. The method is applied for classifying textures in natural images. The analysis of the results shows that the performance of the proposed method is superior to other hybrid methods and the single usage of existing classification methods. [Copyright &y& Elsevier]