Ruxandra Stoean

• PhD
• Professor (Associate) at University of Craiova

Respiratory conditions have been a focal point in recent medical studies. Early detection and timely treatment are crucial factors in improving patient outcomes for any medical condition. Traditionally, doctors diagnose respiratory conditions through an investigation process that involves listening to the patient’s lungs. This study explores the po...

Monitoring heart electrical activity is an effective way of detecting existing and developing conditions. This is usually performed as a non-invasive test using a network of up to 12 sensors (electrodes) on the chest and limbs to create an electrocardiogram (ECG). By visually observing these readings, experienced professionals can make accurate dia...

This paper proposes the use of pre-trained semantic inpainting deep learning architectures to reach a high-fidelity, visually plausible filling content suggestion for the restoration of museum textile objects with traditional motifs. Two state-of-the-art models are selected and their reconstructions are additionally given to an autoencoder trained...

The semantic segmentation for irregularly and not uniformly disposed patterns becomes even more difficult when the occurrence of categories is imbalanced within the images. One example is represented by heavily corroded artefacts in archaeological digs. The current study therefore proposes a weighted loss function within a deep learning architectur...

As solar energy generation has become more and more important for the economies of numerous countries in the last couple of decades, it is highly important to build accurate models for forecasting the amount of green energy that will be produced. Numerous recurrent deep learning approaches, mainly based on long short-term memory (LSTM), are propose...

An effective energy oversight represents a major concern throughout the world, and the problem has become even more stringent recently. The prediction of energy load and consumption depends on various factors such as temperature, plugged load, etc. The machine learning and deep learning (DL) approaches developed in the last decade provide a very hi...

Background Congenital Heart Disease represents the most frequent fetal malformation. The lack of prenatal identification of congenital heart defects can have adverse consequences for the neonate, while a correct prenatal diagnosis of specific cardiac anomalies improves neonatal care neurologic and surgery outcomes. Sonographers perform prenatal dia...

The Schumann Resonances arise from the constructive interference of dozens of near-simultaneous lightning strikes every second, mostly located in the tropics. Characterizing the Schumann Resonance signal variation is a complex task due to the number of variables affecting the electromagnetic composition of the ionosphere and the Earth. We describe...

Fetal echocardiography has become a recommended investigation during pregnancy routine examination in the second semester of fetal development. The heart structures are now sufficiently developed to detect any indication of a possible congenital heart disease. However, the inspection of the key components is not as straightforward for the less expe...

The assessment of the degradation state of an unearthed ancient artefact concerns the identification of the material composition and of the corrosive compounds that are present at the surface. The standard investigation makes use of a combination of invasive and non-invasive techniques and complex devices, while it also relies on the extensive expe...

: The relationship between Schumann resonances and earthquakes was proposed more than 50 years ago; however, the experimental support has not been fully established. A considerable amount of recent studies have focused on the relationship between a single earthquake and the Schumann resonance signal variation around this earthquake, obtaining preli...

Photovoltaic Power is an interesting type of renewable energy, but the intermittency of solar energy resources makes its prediction an challenging task. This article presents the performance of a Hybrid Convolutional - Long short term memory network (CNN-LSTM) architecture in the prediction of photovoltaic generation. The combination was deemed imp...

Schumann resonance (SR) signals result from an electromagnetic wave that propagates along the Earth-ionosphere cavity. This signal is mainly characterized by the amplitude and frequency of the first modes; however, the relation with other variables that affect the Earth-ionosphere cavity is still undiscovered. In this article, this relation is furt...

Swarm intelligence techniques have been created to respond to theoretical and practical global optimization problems. This paper puts forward an enhanced version of the firefly algorithm that corrects the acknowledged drawbacks of the original method, by an explicit exploration mechanism and a chaotic local search strategy. The resulting augmented...

Introduction Over the last decades, a large body of literature has shown that intrapartum clinical digital pelvic estimations of fetal head position, station and progression in the pelvic canal are less accurate, compared with ultrasound (US) scan. Given the increasing evidence regarding the advantages of using US to evaluate the mechanism of labou...

The paper puts forward a convolutional neural network model for multi-output regression, which is trained on images from two distinct microscope types to estimate the concentration of a pair of chemical elements from the surface of archaeological metal objects. The target is to simulate the approximation behaviour of the more complex XRF technology...

The current paper challenges convolutional neural networks to address the computationally undebated task of recognizing four key views in first trimester fetal heart scanning (the aorta, the arches, the atrioventricular flows and the crossing of the great vessels). This is the primary inspection of the heart of a future baby and an early recognitio...

The restoration of archaeological artefacts is naturally utterly important for preserving the cultural heritage. The first step that is undertaken in this process is the chemical analysis of the object, in order to decide the best procedures for its restoration. The gold standard in approximating the concentration of the elements in its composition...

The improvement of combustion processes in industry, especially in the automotive branch, is of great importance to maintain the environmental permitted limits. Carbon monoxide concentration in the exhaust gases can give an insight into the efficiency of the combustion taking place and for this reason, it is important to have sensors that can measu...

The application of echo state networks to time series prediction has provided notable results, favored by their reduced computational cost, since the connection weights require no learning. However, there is a need for general methods that guide the choice of parameters (particularly the reservoir size and ridge regression coefficient), improve the...

Medical data are often tricky to get mined for patterns even by the generally demonstrated successful modern methodologies of deep learning. This paper puts forward such a medical classification task, where patient registers of two of the categories are sometimes hard to be distinguished because of samples showing characteristics of both labels in...

Uncertainty quantification in deep learning models is especially important for the medical applications of this complex and successful type of neural architectures. One popular technique is Monte Carlo dropout that gives a sample output for a record, which can be measured statistically in terms of average probability and variance for each diagnosti...

Application of deep learning (DL) to the field of healthcare is aiding clinicians to make an accurate diagnosis. DL provides reliable results for image processing and sensor interpretation problems most of the time. However, model uncertainty should also be thoroughly quantified. This paper therefore addresses the employment of Monte Carlo dropout...

The paper introduces a novel modality to efficiently tune the convolutional layers of a deep neural network (CNN) and an approach to also rank the importance of the involved hyperparameters. Evolutionary algorithms (EA) offer a flexible solution to this twofold issue, while the expensive simulations of the deep learner with the generated configurat...

Stock price prediction is a popular yet challenging task and deep learning provides the means to conduct the mining for the different patterns that trigger its dynamic movement. In this paper, the task is to predict the close price for 25 companies enlisted at the Bucharest Stock Exchange, from a novel data set introduced herein. Towards this scope...

The present paper puts forward an incipient study that uses a traditional segmentation method based on Zernike moments for extracting significant features from frames of fetal echocardiograms from first trimester color Doppler examinations. A distance based approach is then used on the obtained indicators to classify frames of three given categorie...

There have been important developments in microscopy hardware in the last decades and the number of hospitals to deploy such tools has increased, naturally leading to a very high number of microscopic images that need to be processed. The interest in automated image analysis has accordingly grown and calls for a close collaboration between physicia...

In this paper we introduce a novel methodology for unsupervised analysis of time series, based upon the iterative implementation of a clustering algorithm embedded into the evolution of a recurrent Echo State Network. The main features of the temporal data are captured by the dynamical evolution of the network states, which are then subject to a cl...

This paper aims at assessing spino-cerebellar type 2 ataxia by classifying electrooculography records into registers corresponding to healthy, presymptomatic and ill individuals. The primary used technique is the convolutional neural network applied to the time series of eye movements, called saccades. The problem is exceptionally hard, though, bec...

One prevalent option used in the manufacturing of dental and orthopedic medical implants is titanium, since it is a strong, yet light, biocompatible metal. Nevertheless, possible micro-defects due to earlier chemical treatment can alter its surface morphology and lead to less resistance of the material for implantation. The scope of the present pap...

In order to assemble effective protective coatings against corrosion, electrochemical techniques such as linear potentiometry and cyclic voltammetry were performed on a copper surface in 0.1 mol·L−1 HCl solution containing 0.1% polyvinyl alcohol (PVA) in the absence and presence of silver nanoparticles (nAg/PVA). A recent paradigm was used to disti...

We analyzed 82 patients with colorectal cancer (CRC) [75 patients with mucinous adenocarcinoma (ADK) and seven patients with "signet ring cell" ADK] using multi-cytokeratin (CK) AE1∕AE3 immunohistochemical assay. In order to determine the mucinous nature of some of the lymph node metastases of the mucinous colorectal ADKs studied, Periodic Acid Sch...

In order to design effective protective coatings against corrosion, the polyvinyl alcohol (PVA) as compound and composite with silver nanoparticles (nAg/PVA) were electrodeposited on copper surface employing electrochemical techniques such as linear potentiometry and cyclic voltammetry. A new paradigm was used to distinguish the features of coating...

The Convolutional Neural Network (CNN) approach of deep learning was successfully applied to a field of interest such as metal surface science with modified morphology via corrosion performed in the presence and absence of inhibitors. Thus, given many microscopy images, the artificial intelligence technique learns distinctive features for each clas...

Stock price prediction over time is a problem of practical concern in economics and of scientific interest in financial time series forecasting. The matter also expands toward detecting the variables that play an important role in its behaviour. The current study thus appoints an ARIMA model with regressors to predict the daily return of ten compan...

Many of the current scientific advances in the life sciences have their origin in the intensive use of data for knowledge discovery. In no area this is so clear as in bioinformatics, led by technological breakthroughs in data acquisition technologies. It has been argued that bioinformatics could quickly become the field of research generating the l...

82 patients, who had been diagnosed with colo-rectal adenocarcinoma in our department between 2007 and 2014, were included in our study. Additionally, 31 patients with colo-rectal polyps (20 conventional adenomatous polyps and 11 malignant colo-rectal polyps) were also included in this study. The patients with colo-rectal adenocarcinoma were reeval...

82 patients, who had been diagnosed with colo-rectal adenocarcinoma in our department between 2007 and 2014, were included in our study. Additionally, 31 patients with colo-rectal polyps (20 conventional adenomatous polyps and 11 malignant colo-rectal polyps) were also included in this study. The patients with colo-rectal adenocarcinoma were reeval...

The paper presents an efficient design and implementation of a convolutional neu-ral network on an FPGA device. The aim is not only theoretical but also practical, since the solution will be used in a medical clinic dealing with SpinoCerebellar Ataxia type 2 as part of a larger project. Hence, the current work targets both high learning capabilitie...

A precise estimation of patient length of stay is important for systematically managing both hospital unit resources (medication, equipment, beds) and the distribution of personnel. This is true for hospitalization following any disease, however the particularities of each trigger a different observation/recovery period. The current study investiga...

The paper addresses the medical challenge of interpreting histopathological slides through expert-independent automated learning with implicit feature determination and direct grading establishment. Deep convolutional neural networks model the image collection and are able to give a timely and accurate support for pathologists, who are more than of...

The paper puts forward a new data set comprising 357 histopathological image samples obtained from colon tissues and distinguished into four cancer grades. At the same time, it proposes an automatic methodology for extracting knowledge from these images and discriminating between the disease stages on its base. The approach identifies the glands an...

A genetic algorithm adjusts parameters of image segmentation techniques to be transposed from manually annotated cropped images to perform on the complete contour-blank histopathological pictures.

The paper puts forward an ensemble of state-of-the-art classifiers – support vector machines, neural networks and decision trees – to estimate the length of stay after surgery in patients diagnosed with colorectal cancer. The three paradigms are brought together in order to achieve both a more accurate prediction through a voting scheme and transpa...

The response of a computational system to support medical diagnosis should simultaneously be accurate, comprehensible, flexible and prompt in order to be qualified as a reliable second opinion. Based on the above characteristics, the current paper examines the behaviour of two evolutionary algorithms that discover prototypes for each possible diagn...

Although neural networks and support vector machines (SVMs) are the traditional predictors for the classification of complex problems, these opaque paradigms cannot explain the logic behind the discrimination process. Therefore, within the quite unexplored area of evolutionary algorithms opening the SVM decision black box, the paper appoints a coop...

The natural evolution of individuals implies adaptation to the environment, part of which consists of other living beings, in particular different groups or species. From this viewpoint, evolution is actually coevolution. Coevolution can be competitive, cooperative or both. Similarly, in the evolutionary computational area, EAs have been also exten...

Every aspect of our daily life implies a search for the best possible action and the optimal choice for that act. Most standard optimization methods require the fulfilment of certain constraints, imply convergence issues or use single point movement. Among them, EAs represent a flexible and adaptable alternative, with classes of methods based on pr...

As mentioned in the introductory chapter, for the classification tasks we aim to evolve thresholds for the attributes of the given data training examples. Such vectors of thresholds will be constructed for each class of the problem and we will further refer to them as prototypes. They will efficiently represent the class distributions and therefore...

The kernel-based methodology of SVMs [Vapnik and Chervonenkis, 1974], [Vapnik, 1995a] has been established as a top ranking approach for supervised learning within both the theoretical and red practical research environments. This very performing technique suffers nevertheless from the curse of an opaque engine [Huysmans et al, 2006], which is unde...

Even if SVMs are one of the most reliable classifiers for real-world tasks when it comes to accurate prediction, their weak point still lies in the opacity behind their resulting discrimination [Huysmans et al, 2006]. As we have mentioned before, there aremany available implementations that offer the possibility to also extract the coefficients of...

We have reached the end of the book. Have we got any answers to the question we raised in the introductory chapter? Having presented the several options for classification – SVMs alone, asingle EAs and hybridization at two stages of learning – what choice proved to be more advantageous, taking into consideration prediction accuracy, comprehensibili...

In the current study, parameter tuning is performed for two evolutionary optimization techniques, Covariance Matrix Adaptation Evolution Strategy and Topological Species Conservation. They are applied for three multimodal benchmark functions with various properties and several outputs are considered. A data set with input parameters and metaheurist...

Machine learning support for medical decision making is truly helpful only when it meets two conditions: high prediction accuracy and a good expla-nation of how the diagnosis was reached. Support vector machines (SVMs) successfully achieve the first target due to a kernel-based engine; evolutionary algorithms (EAs) can greatly accomplish the second...

The paper presents a review of current evolutionary algorithms for feature ranking in data mining tasks involving automated learning. This issue is highly important as real-world problems commonly suffer from the curse of dimensionality. By weighting the significance of each attribute from a data set, the less influential indicators can be disposed...

The performance of niching based or related evolutionary algorithms clearly depends on problem properties as e.g. the number of local optima of a problem. We assume there must be more such properties currently not taken into account and, following from practical experience, suggest two more, namely basin size contrast (BSC), the size relation of th...

This paper presents an automatic tool capable to learn from a patients data set with 24 medical indicators characterizing each sample and to subsequently use the acquired knowledge to differentiate between five degrees of liver fibrosis. The indicators represent clinical observations and the liver stiffness provided by the new, non-invasive procedu...

Objective: Hepatic fibrosis, the principal pointer to the development of a liver disease within chronic hepatitis C, can be measured through several stages. The correct evaluation of its degree, based on recent different non-invasive procedures, is of current major concern. The latest methodology for assessing it is the Fibroscan and the effect of...

Any evolutionary technique for multimodal optimization must answer two crucial questions in order to guarantee some success on a given task: How to most unboundedly distinguish between the different attraction basins and how to most accurately safeguard the consequently discovered solutions. This paper thus aims to present a novel technique that in...

Present paper brings together two novel evolutionary techniques designed for clas-sification and applied for the differentiation among five possible degrees of liver fibrosis within chronic hepatitis C. A purely evolutionary method -the cooperative coevolutionary classifier -endowed with a hill climbing algorithm for the selection of influential at...

Solving multimodal optimization tasks (problems with multiple glo-bal/local optimal solutions) by the state-of-the-art evolutionary algorithms (EAs) presumes separation of a population of individuals into subpopulations, each connected to a different optimum, with the aim of maintaining diversity for a longer period of time. Instead of using the ty...

The paper presents a novel evolutionary technique constructed as an alternative of the standard support vector machines architecture. The approach adopts the learning strategy of the latter but aims to simplify and generalize its training, by offering a transparent substitute to the initial black-box. Contrary to the canonical technique, the evolut...

Individuals encoding potential rules to model an actual partition of samples into categories may be evolved by means of several well-known evolutionary classification techniques. Nevertheless, since a canonical evolutionary algorithm progresses towards one (global or local) optimum, some special construction or certain additional method are designe...

Support vector machines represent a state-of-the-art paradigm, which has nevertheless been tackled by a number of other approaches in view of the development of a superior hybridized technique. It is also the proposal of present chapter to bring support vector machines together with evolutionary computation, with the aim to offer a simplified solvi...

The paper presents a novel, combined methodology to target parameter tuning. It uses Latin hypersquare sampling to generate a diverse, large set of configurations for the variables to be set. These serve as input for the metaheuristic to be tuned and a complete data set, with the parameter values and the success rate obtained by the algorithm, is f...

The present paper introduces a new evolutionary technique for multimodal real-valued optimization which uses a clustering method for separating the individuals within a population into species that are each connected to different optima from the search space. It is applied for a set of benchmark functions both for uni- and multimodal optimization a...

When using an evolutionary algorithm on an unknown problem, properties like the number of global/local optima must be guessed for properly picking an algorithm and its parameters. It is the aim of current paper to put forward an EA-based method for real-valued optimization to provide an estimate on the number of optima a function exhibits, or at le...

Purpose – The purpose of this paper is to verify whether an evolutionary model outperforms logistic regression in determining the institutional placement decisions made by a London social service department panel. Design/methodology/approach – Genetic chromodynamics models an algorithm within the Michigan evolutionary classifier. Hence multiple cla...

Evolutionary algorithms for multimodal optimization are usually radius dependent in subpopulation formation and dynamics. The appropriate setting of a value to cover such a fluctuant threshold (as a result of variously shaped attraction basins) strongly relies upon the knowledge on the problem at hand or the extensive skill in manual parameter tuni...

A data mining field with daily, and sometimes even vital, practical applications, classification has been addressed by many powerful paradigms, among which evolutionary algorithms (EAs) play a successful role. Nevertheless, as evolutionary computation (EC) progresses, there appear new possibilities of developing simpler and yet robust classificatio...

Within the present paper, we put forward a novel hybridization between support vector machines and evolutionary algorithms. Evolutionary support vector machines consider the classification task as in support vector machines but use an evolutionary algorithm to solve the optimization problem of determining the decision function. They can explicitly...

The present paper investigates the hybridization of two well- known multimodal optimization methods, i.e. species con- servation and multinational algorithms. The topological species conservation algorithm embraces the vision of the existence of subpopulations around seeds (the best local in- dividuals) and the preservation of these dominating indi...

The aim of this paper is to validate the new paradigm of evolutionary support vector machines (ESVMs) for binary classification also through an application to a real-world problem, i.e. the diagnosis of diabetes mellitus. ESVMs were developed through hybridization between the strong learning paradigm of support vector machines (SVMs) and the optimi...

Evolutionary support vector machines (ESVMs) are a novel technique that assimilates the learning engine of the state-of-the-art support vector machines (SVMs) but evolves the coefficients of the decision function by means of evo- lutionary algorithms (EAs). The new method has accom- plished the purpose for which it has been initially devel- oped, t...

A new learning technique based on cooperative coevo- lution is proposed for tackling classification problems. For each possible outcome of the classification task, a popu- lation of if-then rules, all having that certain class as the conclusion part, is evolved. Cooperation between rules ap- pears in the evaluation stage, when complete sets of rule...

We propose a novel learning technique for classification as result of the hybridization between support vector machines and evolutionary algorithms. Evolutionary support vector machines consider the classification task as in support vector machines but use evolutionary algorithms to solve the optimization problem of determining the decision functio...

Present paper adresses the famous machine learning paradigm, called support vector machines, from the viewpoint of evolutionary computation. Namely, the constrained optimization problem within support vector machines is solved through an evolutionary algorithm, for the sake of simplicity. The new approach has been so far applied solely to the case...

Support vector machines are a modern and very efficient learning heuristic. However, their internal engine relies on not very easy or common mathematical concepts. The paper presents a newly developed simpler de-sign of the engine, built through the means of evolutionary computation, in the context of nonlinear support vector machines. Experiments...

An evolutionary algorithm based on cooperative coevolution is applied to a classification problem, the Pima Indian diabetes diagnosis problem. Previous cooperative coevolution algorithms were developed for function optimization [1], optimizing agents behaviour [2] or modelling the behaviour of a robot in an unknown environment [3]. The aim of this...

Evolutionary support vector machines represent a new learning technique that we recently developed as a hybridization between support vector machines and evolutionary algorithms, regarding the discovery of the optimal decision function within the former. The new approach has proven to be successful as binary classification problems have been concer...

A new radii-based evolutionary algorithm (EA) designed for multimodal optimization problems is proposed. The approach can be placed within the genetic chromodynamics framework and related to other EAs with local interaction, e.g. using species formation or clearing procedures. The underlying motivation for modifying the original algorithm was to pr...

Accurate spam filters are of high necessity in present days as the high amount of commercial mail entering accounts has become a real threat to everyone, from causing personal computers to crash to costing big companies billions of dollars annually because of employees loss of productivity. Moreover, lately, spam also carries viruses along. Current...