Median radial basis function neural network

Considering radial basis function neural network for effective solution generation in metaheuristic algorithms

In many engineering optimization problems, the number of function evaluations is severely limited by the time or cost constraints. These limitations present a significant challenge in the field of global optimization, because existing metaheuristic methods typically require a substantial number of function evaluations to find optimal solutions. This paper presents a new metaheuristic optimization algorithm that considers the information obtained by a radial basis function neural network (RBFNN) in terms of the objective function for guiding the search process. Initially, the algorithm uses the maximum design approach to strategically distribute a set of solutions across the entire search space. It then enters a cycle in which the RBFNN models the objective function values from the current solutions. The algorithm identifies and uses key neurons in the hidden layer that correspond to the highest objective function values to generate new solutions. The centroids and standard deviations of these neurons guide the sampling process, which continues until the desired number of solutions is reached. By focusing on the areas of the search space that yield high objective function values, the algorithm avoids exhaustive solution evaluations and significantly reduces the number of function evaluations. The effectiveness of the method is demonstrated through a comparison with popular metaheuristic algorithms across several test functions, where it consistently outperforms existing techniques, delivers higher-quality solutions, and improves convergence rates.

Polarimetric Imaging for Robot Perception: A Review

In recent years, the integration of polarimetric imaging into robotic perception systems has increased significantly, driven by the accessibility of affordable polarimetric sensors. This technology complements traditional color imaging by capturing and analyzing the polarization characteristics of light. This additional information provides robots with valuable insights into object shape, material composition, and other properties, ultimately enabling more robust manipulation tasks. This review aims to provide a comprehensive analysis of the principles behind polarimetric imaging and its diverse applications within the field of robotic perception. By exploiting the polarization state of light, polarimetric imaging offers promising solutions to three key challenges in robot vision: Surface segmentation; depth estimation through polarization patterns; and 3D reconstruction using polarimetric data. This review emphasizes the practical value of polarimetric imaging in robotics by demonstrating its effectiveness in addressing real-world challenges. We then explore potential applications of this technology not only within the core robotics field but also in related areas. Through a comparative analysis, our goal is to elucidate the strengths and limitations of polarimetric imaging techniques. This analysis will contribute to a deeper understanding of its broad applicability across various domains within and beyond robotics.

KeratoEL: Detection of keratoconus using corneal parameters with ensemble learning

Background and Aims

Keratoconus is a progressive eye condition in which the normally round cornea thins and bulges outwards into a cone shape. This irregular shape causes light to scatter in multiple directions as it enters the eye, leading to distorted vision, increased sensitivity to light and frequent changes in the prescription of glasses or contact lenses. Detecting keratoconus at an early stage is not only difficult but also challenging.

Methods

The study has proposed an ensemble-based machine learning (ML) technique named KeratoEL to detect keratoconus at an early stage. The proposed KeratoEL model combines the basic machine learning algorithms, namely support vector machine (SVM), decision tree (DT), random forest (RF) and artificial neural network (ANN). Before employing the ML model for keratoconus detection, the data set is first preprocessed manually by eliminating some features that don't contribute any significant value to predict the exact class. Moreover, the output features are labelled into three different classes and Extra Trees Classifier is used to find out the important features. Then, the features are sorted in descending order and top 45, 30, and 15 features are taken as input datasets against the output. Finally, different machine learning models are tested using the input datasets and performance metrics are measured.

Results

The proposed model obtains 98.0%, 98.9% and 99.8% accuracy for top 45, 30, and 15 number of features respectively. Overall experimental results show that the proposed ensemble model outperforms the existing machine learning models.

Conclusion

The proposed KeratoEL model effectively detects keratoconus at an early stage by combining SVM, DT, RF, and ANN algorithms, demonstrating superior performance over existing models. These results underscore the potential of the KeratoEL ensemble approach in enhancing early detection and treatment of keratoconus.

Using radial basis artificial neural networks to predict radiation hazard indices in geological materials

The estimation of exposures to humans from the various sources of radiation is important. Radiation hazard indices are computed using procedures described in the literature for evaluating the combined effects of the activity concentrations of primordial radionuclides, namely, 238U, 232Th, and 40 K. The computed indices are then compared to the allowed limits defined by International Radiation Protection Organizations to determine any radiation hazard associated with the geological materials. In this paper, four distinct radial basis function artificial neural network (RBF-ANN) models were developed to predict radiation hazard indices, namely, external gamma dose rates, annual effective dose, radium equivalent activity, and external hazard index. To make RBF-ANN models, 348 different geological materials' gamma spectrometry data were acquired from the literature. Radiation hazards indices predicted from each RBF-ANN model were compared to the radiation hazards calculated using gamma spectrum analysis. The predicted hazard indices values of each RBF-ANN model were found to precisely align with the calculated values. To validate the accuracy and the adaptability of each RBF-ANN model, statistical tests (determination coefficient (R2), relative absolute error (RAE), root mean square error (RMSE), Nash-Sutcliffe Efficiency (NSE)), and significance tests (F-test and Student's t-test) were performed to analyze the relationship between calculated and predicted hazard indices. Low RAE and RMSE values as well as high R2, NSE, and p-values greater than 0.95, 0.71, and 0.05, respectively, were found for RBF-ANN models. The statistical tests' results show that all RBF-ANN models created exhibit precise performance, indicating their applicability and efficiency in forecasting the radiation hazard indices of geological materials. All the RBF-ANN models can be used to predict radiation hazard indices of geological materials quite efficiently, according to the performance level attained.

Parkinson's disease detection from EEG signal employing autoencoder and RBFNN-based hybrid deep learning framework utilizing power spectral density

Objective

Early detection of Parkinson's disease (PD) is essential for halting its progression, yet challenges remain in leveraging deep learning for accurate identification. This study aims to overcome these obstacles by introducing a hybrid deep learning approach that enhances PD detection through a combination of autoencoder (AE) and radial basis function neural network (RBFNN).

Methods

The proposed method analyzes the power spectral density (PSD) of preprocessed electroencephalography (EEG) signals, with artifacts removed, to assess energy distribution across EEG sub-bands. AEs are employed to extract features from reconstructed signals, which are subsequently classified by an RBFNN. The approach is validated on UC SanDiego's EEG dataset, consisting of 31 subjects and 93 minutes of recordings.

Results

The hybrid model demonstrates promising performance, achieving a classification accuracy of 99%. The improved accuracy is attributed to advanced feature selection techniques, robust data preprocessing, and the integration of AEs with RBFNN, setting a new benchmark in PD detection frameworks.

Conclusion

This study highlights the efficacy of the hybrid deep learning framework in detecting PD, particularly emphasizing the importance of using multiple EEG channels and advanced preprocessing techniques. The results underscore the potential of this approach for practical clinical applications, offering a reliable solution for early and accurate PD detection.

Deep Mixture Generative Autoencoders

Variational autoencoders (VAEs) are one of the most popular unsupervised generative models that rely on learning latent representations of data. In this article, we extend the classical concept of Gaussian mixtures into the deep variational framework by proposing a mixture of VAEs (MVAE). Each component in the MVAE model is implemented by a variational encoder and has an associated subdecoder. The separation between the latent spaces modeled by different encoders is enforced using the d -variable Hilbert-Schmidt independence criterion (dHSIC). Each component would capture different data variational features. We also propose a mechanism for finding the appropriate number of VAE components for a given task, leading to an optimal architecture. The differentiable categorical Gumbel-softmax distribution is used in order to generate dropout masking parameters within the end-to-end backpropagation training framework. Extensive experiments show that the proposed MVAE model can learn a rich latent data representation and is able to discover additional underlying data representation factors.

A Method for Separating Multisource Partial Discharges in a Substation Based on Selected Bispectra of UHF Signals

A method for separating multisource partial discharges (PDs) in a substation is proposed based on selected bispectra of ultrahigh frequency (UHF) electromagnetic waves. Bispectra are sensitive to Gaussian noises and processes of symmetrical distribution. The phase information contained in bispectra can be useful and important for further signal processing. Bifrequencies where Fisher-like class separability measures between signals’ bispectra achieve their maximums are selected as characteristic parameters of the signals. Then, the selected bispectra are utilized for training the radial basis neural network to separate PD UHF signals in a substation. The method is used to analyze simulated UHF signals mixed with Gaussian white noise and frequency-fixed interference, and to separate PD UHF signals that are collected in a 500 kV substation. In order to prove the validity of the proposed separation method, the localization results are compared with the results calculated by time delay sequence, and the proposed separating algorithm is verified in the interference circumstances of a substation. However, the exact location of PD sources cannot be calculated according to the time delay sequence when the PD sources in a substation are close to each other or there are fewer than four antennas for receiving signals.

Scalable Bayesian preference learning for crowds

We propose a scalable Bayesian preference learning method for jointly predicting the preferences of individuals as well as the consensus of a crowd from pairwise labels. Peoples’ opinions often differ greatly, making it difficult to predict their preferences from small amounts of personal data. Individual biases also make it harder to infer the consensus of a crowd when there are few labels per item. We address these challenges by combining matrix factorisation with Gaussian processes, using a Bayesian approach to account for uncertainty arising from noisy and sparse data. Our method exploits input features, such as text embeddings and user metadata, to predict preferences for new items and users that are not in the training set. As previous solutions based on Gaussian processes do not scale to large numbers of users, items or pairwise labels, we propose a stochastic variational inference approach that limits computational and memory costs. Our experiments on a recommendation task show that our method is competitive with previous approaches despite our scalable inference approximation. We demonstrate the method’s scalability on a natural language processing task with thousands of users and items, and show improvements over the state of the art on this task. We make our software publicly available for future work (https://github.com/UKPLab/tacl2018-preference-convincing/tree/crowdGPPL).

R2BN: An Adaptive Model for Keystroke-Dynamics-Based Educational Level Classification

Over the past decade, keystroke-based pattern recognition techniques, as a forensic tool for behavioral biometrics, have gained increasing attention. Although a number of machine learning-based approaches have been proposed, they are limited in terms of their capability to recognize and profile a set of an individual's characteristics. In addition, up to today, their focus was primarily gender and age, which seem to be more appropriate for commercial applications (such as developing commercial software), leaving out from research other characteristics, such as the educational level. Educational level is an acquired user characteristic, which can improve targeted advertising, as well as provide valuable information in a digital forensic investigation, when it is known. In this context, this paper proposes a novel machine learning model, the randomized radial basis function network, which recognizes and profiles the educational level of an individual who stands behind the keyboard. The performance of the proposed model is evaluated by using the empirical data obtained by recording volunteers' keystrokes during their daily usage of a computer. Its performance is also compared with other well-referenced machine learning models using our keystroke dynamic datasets. Although the proposed model achieves high accuracy in educational level prediction of an unknown user, it suffers from high computational cost. For this reason, we examine ways to reduce the time that is needed to build our model, including the use of a novel data condensation method, and discuss the tradeoff between an accurate and a fast prediction. To the best of our knowledge, this is the first model in the literature that predicts the educational level of an individual based on the keystroke dynamics information only.

KeratoDetect: Keratoconus Detection Algorithm Using Convolutional Neural Networks

Keratoconus (KTC) is a noninflammatory disorder characterized by progressive thinning, corneal deformation, and scarring of the cornea. The pathological mechanisms of this condition have been investigated for a long time. In recent years, this disease has come to the attention of many research centers because the number of people diagnosed with keratoconus is on the rise. In this context, solutions that facilitate both the diagnostic and treatment options are quickly needed. The main contribution of this paper is the implementation of an algorithm that is able to determine whether an eye is affected or not by keratoconus. The KeratoDetect algorithm analyzes the corneal topography of the eye using a convolutional neural network (CNN) that is able to extract and learn the features of a keratoconus eye. The results show that the KeratoDetect algorithm ensures a high level of performance, obtaining an accuracy of 99.33% on the data test set. KeratoDetect can assist the ophthalmologist in rapid screening of its patients, thus reducing diagnostic errors and facilitating treatment.

Multicolumn RBF Network

This paper proposes the multicolumn RBF network (MCRN) as a method to improve the accuracy and speed of a traditional radial basis function network (RBFN). The RBFN, as a fully connected artificial neural network (ANN), suffers from costly kernel inner-product calculations due to the use of many instances as the centers of hidden units. This issue is not critical for small datasets, as adding more hidden units will not burden the computation time. However, for larger datasets, the RBFN requires many hidden units with several kernel computations to generalize the problem. The MCRN mechanism is constructed based on dividing a dataset into smaller subsets using the k-d tree algorithm. resultant subsets are considered as separate training datasets to train individual RBFNs. Those small RBFNs are stacked in parallel and bulged into the MCRN structure during testing. The MCRN is considered as a well-developed and easy-to-use parallel structure, because each individual ANN has been trained on its own subsets and is completely separate from the other ANNs. This parallelized structure reduces the testing time compared with that of a single but larger RBFN, which cannot be easily parallelized due to its fully connected structure. Small informative subsets provide the MCRN with a regional experience to specify the problem instead of generalizing it. The MCRN has been tested on many benchmark datasets and has shown better accuracy and great improvements in training and testing times compared with a single RBFN. The MCRN also shows good results compared with those of some machine learning techniques, such as the support vector machine and k-nearest neighbors.

Plantar fascia segmentation and thickness estimation in ultrasound images

Ultrasound (US) imaging offers significant potential in diagnosis of plantar fascia (PF) injury and monitoring treatment. In particular US imaging has been shown to be reliable in foot and ankle assessment and offers a real-time effective imaging technique that is able to reliably confirm structural changes, such as thickening, and identify changes in the internal echo structure associated with diseased or damaged tissue. Despite the advantages of US imaging, images are difficult to interpret during medical assessment. This is partly due to the size and position of the PF in relation to the adjacent tissues. It is therefore a requirement to devise a system that allows better and easier interpretation of PF ultrasound images during diagnosis. This study proposes an automatic segmentation approach which for the first time extracts ultrasound data to estimate size across three sections of the PF (rearfoot, midfoot and forefoot). This segmentation method uses artificial neural network module (ANN) in order to classify small overlapping patches as belonging or not-belonging to the region of interest (ROI) of the PF tissue. Features ranking and selection techniques were performed as a post-processing step for features extraction to reduce the dimension and number of the extracted features. The trained ANN classifies the image overlapping patches into PF and non-PF tissue, and then it is used to segment the desired PF region. The PF thickness was calculated using two different methods: distance transformation and area-length calculation algorithms. This new approach is capable of accurately segmenting the PF region, differentiating it from surrounding tissues and estimating its thickness.

Evaluation of machine learning classifiers in keratoconus detection from orbscan II examinations

PURPOSE

To evaluate the performance of support vector machine, multi-layer perceptron and radial basis function neural network as auxiliary tools to identify keratoconus from Orbscan II maps.

METHODS

A total of 318 maps were selected and classified into four categories: normal (n = 172), astigmatism (n = 89), keratoconus (n = 46) and photorefractive keratectomy (n = 11). For each map, 11 attributes were obtained or calculated from data provided by the Orbscan II. Ten-fold cross-validation was used to train and test the classifiers. Besides accuracy, sensitivity and specificity, receiver operating characteristic (ROC) curves for each classifier were generated, and the areas under the curves were calculated.

RESULTS

The three selected classifiers provided a good performance, and there were no differences between their performances. The area under the ROC curve of the support vector machine, multi-layer perceptron and radial basis function neural network were significantly larger than those for all individual Orbscan II attributes evaluated (p < 0.05).

CONCLUSION

Overall, the results suggest that using a support vector machine, multi-layer perceptron classifiers and radial basis function neural network, these classifiers, trained on Orbscan II data, could represent useful techniques for keratoconus detection.

Minimum class variance support vector machines

In this paper, a modified class of support vector machines (SVMs) inspired from the optimization of Fisher's discriminant ratio is presented, the so-called minimum class variance SVMs (MCVSVMs). The MCVSVMs optimization problem is solved in cases in which the training set contains less samples that the dimensionality of the training vectors using dimensionality reduction through principal component analysis (PCA). Afterward, the MCVSVMs are extended in order to find nonlinear decision surfaces by solving the optimization problem in arbitrary Hilbert spaces defined by Mercer's kernels. In that case, it is shown that, under kernel PCA, the nonlinear optimization problem is transformed into an equivalent linear MCVSVMs problem. The effectiveness of the proposed approach is demonstrated by comparing it with the standard SVMs and other classifiers, like kernel Fisher discriminant analysis in facial image characterization problems like gender determination, eyeglass, and neutral facial expression detection.

Variational learning for Gaussian mixture models

This paper proposes a joint maximum likelihood and Bayesian methodology for estimating Gaussian mixture models. In Bayesian inference, the distributions of parameters are modeled, characterized by hyperparameters. In the case of Gaussian mixtures, the distributions of parameters are considered as Gaussian for the mean, Wishart for the covariance, and Dirichlet for the mixing probability. The learning task consists of estimating the hyperparameters characterizing these distributions. The integration in the parameter space is decoupled using an unsupervised variational methodology entitled variational expectation-maximization (VEM). This paper introduces a hyperparameter initialization procedure for the training algorithm. In the first stage, distributions of parameters resulting from successive runs of the expectation-maximization algorithm are formed. Afterward, maximum-likelihood estimators are applied to find appropriate initial values for the hyperparameters. The proposed initialization provides faster convergence, more accurate hyperparameter estimates, and better generalization for the VEM training algorithm. The proposed methodology is applied in blind signal detection and in color image segmentation.

A cost-function approach to rival penalized competitive learning (RPCL)

Rival penalized competitive learning (RPCL) has been shown to be a useful tool for clustering on a set of sample data in which the number of clusters is unknown. However, the RPCL algorithm was proposed heuristically and is still in lack of a mathematical theory to describe its convergence behavior. In order to solve the convergence problem, we investigate it via a cost-function approach. By theoretical analysis, we prove that a general form of RPCL, called distance-sensitive RPCL (DSRPCL), is associated with the minimization of a cost function on the weight vectors of a competitive learning network. As a DSRPCL process decreases the cost to a local minimum, a number of weight vectors eventually fall into a hypersphere surrounding the sample data, while the other weight vectors diverge to infinity. Moreover, it is shown by the theoretical analysis and simulation experiments that if the cost reduces into the global minimum, a correct number of weight vectors is automatically selected and located around the centers of the actual clusters, respectively. Finally, we apply the DSRPCL algorithms to unsupervised color image segmentation and classification of the wine data.

Digital image processing techniques for the detection and removal of cracks in digitized paintings

An integrated methodology for the detection and removal of cracks on digitized paintings is presented in this paper. The cracks are detected by thresholding the output of the morphological top-hat transform. Afterward, the thin dark brush strokes which have been misidentified as cracks are removed using either a median radial basis function neural network on hue and saturation data or a semi-automatic procedure based on region growing. Finally, crack filling using order statistics filters or controlled anisotropic diffusion is performed. The methodology has been shown to perform very well on digitized paintings suffering from cracks.

High-Speed Face Recognition Based on Discrete Cosine Transform and RBF Neural Networks

In this paper, an efficient method for high-speed face recognition based on the discrete cosine transform (DCT), the Fisher's linear discriminant (FLD) and radial basis function (RBF) neural networks is presented. First, the dimensionality of the original face image is reduced by using the DCT and the large area illumination variations are alleviated by discarding the first few low-frequency DCT coefficients. Next, the truncated DCT coefficient vectors are clustered using the proposed clustering algorithm. This process makes the subsequent FLD more efficient. After implementing the FLD, the most discriminating and invariant facial features are maintained and the training samples are clustered well. As a consequence, further parameter estimation for the RBF neural networks is fulfilled easily which facilitates fast training in the RBF neural networks. Simulation results show that the proposed system achieves excellent performance with high training and recognition speed, high recognition rate as well as very good illumination robustness.

Analysis of augmented-input-layer RBFNN

In this paper we present and analyze a new structure for designing a radial basis function neural network (RBFNN). In the training phase, input layer of RBFNN is augmented with desired output vector. Generalization phase involves the following steps: (1) identify the cluster to which a previously unseen input vector belongs; (2) augment the input layer with an average of the targets of the input vectors in the identified cluster; and (3) use the augmented network to estimate the unknown target. It is shown that, under some reasonable assumptions, the generalization error function admits an upper bound in terms of the quantization errors minimized when determining the centers of the proposed method over the training set and the difference between training samples and generalization samples in a deterministic setting. When the difference between the training and generalization samples goes to zero, the upper bound can be made arbitrarily small by increasing the number of hidden neurons. Computer simulations verified the effectiveness of the proposed method.

Best harmony, unified RPCL and automated model selection for unsupervised and supervised learning on Gaussian mixtures, three-layer nets and ME-RBF-SVM models

After introducing the fundamentals of BYY system and harmony learning, which has been developed in past several years as a unified statistical framework for parameter learning, regularization and model selection, we systematically discuss this BYY harmony learning on systems with discrete inner-representations. First, we shown that one special case leads to unsupervised learning on Gaussian mixture. We show how harmony learning not only leads us to the EM algorithm for maximum likelihood (ML) learning and the corresponding extended KMEAN algorithms for Mahalanobis clustering with criteria for selecting the number of Gaussians or clusters, but also provides us two new regularization techniques and a unified scheme that includes the previous rival penalized competitive learning (RPCL) as well as its various variants and extensions that performs model selection automatically during parameter learning. Moreover, as a by-product, we also get a new approach for determining a set of 'supporting vectors' for Parzen window density estimation. Second, we shown that other special cases lead to three typical supervised learning models with several new results. On three layer net, we get (i) a new regularized ML learning, (ii) a new criterion for selecting the number of hidden units, and (iii) a family of EM-like algorithms that combines harmony learning with new techniques of regularization. On the original and alternative models of mixture-of-expert (ME) as well as radial basis function (RBF) nets, we get not only a new type of criteria for selecting the number of experts or basis functions but also a new type of the EM-like algorithms that combines regularization techniques and RPCL learning for parameter learning with either least complexity nature on the original ME model or automated model selection on the alternative ME model and RBF nets. Moreover, all the results for the alternative ME model are also applied to other two popular nonparametric statistical approaches, namely kernel regression and supporting vector machine. Particularly, not only we get an easily implemented approach for determining the smoothing parameter in kernel regression, but also we get an alternative approach for deciding the set of supporting vectors in supporting vector machine.

Prediction and tracking of moving objects in image sequences

We employ a prediction model for moving object velocity and location estimation derived from Bayesian theory. The optical flow of a certain moving object depends on the history of its previous values. A joint optical flow estimation and moving object segmentation algorithm is used for the initialization of the tracking algorithm. The segmentation of the moving objects is determined by appropriately classifying the unlabeled and the occluding regions. Segmentation and optical flow tracking is used for predicting future frames.

Object classification in 3-D images using alpha-trimmed mean radial basis function network

We propose a pattern classification based approach for simultaneous three-dimensional (3-D) object modeling and segmentation in image volumes. The 3-D objects are described as a set of overlapping ellipsoids. The segmentation relies on the geometrical model and graylevel statistics. The characteristic parameters of the ellipsoids and of the graylevel statistics are embedded in a radial basis function (RBF) network and they are found by means of unsupervised training. A new robust training algorithm for RBF networks based on alpha-trimmed mean statistics is employed in this study. The extension of the Hough transform algorithm in the 3-D space by employing a spherical coordinate system is used for ellipsoidal center estimation. We study the performance of the proposed algorithm and we present results when segmenting a stack of microscopy images.

Optical flow estimation and moving object segmentation based on median radial basis function network

Various approaches have been proposed for simultaneous optical flow estimation and segmentation in image sequences. In this study, the moving scene is decomposed into different regions with respect to their motion, by means of a pattern recognition scheme. The inputs of the proposed scheme are the feature vectors representing still image and motion information. Each class corresponds to a moving object. The classifier employed is the median radial basis function (MRBF) neural network. An error criterion function derived from the probability estimation theory and expressed as a function of the moving scene model is used as the cost function. Each basis function is activated by a certain image region. Marginal median and median of the absolute deviations from the median (MAD) estimators are employed for estimating the basis function parameters. The image regions associated with the basis functions are merged by the output units in order to identify moving objects.