Papers by Antonio Quintero-Rincón
Inteligencia Artificial, 2025
Breast cancer is a tumor that begins to grow in the milk ducts or lobules and can become lethal i... more Breast cancer is a tumor that begins to grow in the milk ducts or lobules and can become lethal if treatment is not administered in time. According to the World Health Organization (WHO), there were approximately 2.3 million cases of breast cancer in 2020. Furthermore, breast cancer can affect anyone, particularly women over 50 years old. Therefore, it is crucial to have early diagnostic techniques. We propose a novel method based on Cohen's d for feature selection in this context. Cohen's d is a statistical concept that quantifies the strength of the relationship between two populations on a numeric scale. The central idea is to utilize Cohen's d effect size as a feature selector to reduce the dimensionality of the data and enhance the predictors through a Machine Learning (ML) classifier model for diagnosing breast cancer. For experimental purposes, the Breast Cancer Wisconsin database was used. This proposed feature selector is compared with two classical methods: Learning Vector Quantization (LVQ) and Recursive Feature Elimination (RFE). A random evaluation of the features of each selector was conducted 100 times using a Support Vector Machine (SVM) classifier, resulting in the following average outcomes: Cohen's d based feature selector showed 0.96 sensitivity and 0.97 specificity, RFE based feature selector exhibited 0.95 sensitivity and 0.98 specificity, and LVQ based feature selector demonstrated 0.91 sensitivity and 0.96 specificity. These promising results indicate that the proposed methodology utilizing Cohen's d may be a valuable feature selector and sheds light on the long-standing research into breast cancer detection.
PLoS ONE, 2025
Despite tremendous efforts devoted to the area, image texture analysis is still an open research ... more Despite tremendous efforts devoted to the area, image texture analysis is still an open research field. This paper presents an algorithm and experimental results demonstrating the feasibility of developing automated tools to detect abnormal X-ray images based on tissue attenuation. Specifically, this work proposes using the variability characterised by singular values and conditional indices extracted from the singular value decomposition (SVD) as image texture features. In addition, the paper introduces a "tuning weight" parameter to consider the variability of the X-ray attenuation in tissues affected by pathologies. This weight is estimated using the coefficient of variation of the minimum covariance determinant from the bandwidth yielded by the non-parametric distribution of variance-decomposition proportions of the SVD. When multiplied by the two features (singular values and conditional indices), this single parameter acts as a tuning weight, reducing misclassification and improving the classic performance metrics, such as true positive rate, false negative rate, positive predictive values, false discovery rate, areaunder-curve, accuracy rate, and total cost. The proposed method implements an ensemble bagged trees classification model to classify X-ray chest images as COVID-19, viral pneumonia, lung opacity, or normal. It was tested using a challenging, imbalanced chest X-ray public dataset. The results show an accuracy of 88% without applying the tuning weight and 99% with its application. The proposed method outperforms state-of-the-art methods, as attested by all performance metrics.
Science des données, 2024
La somnolence des conducteurs est une cause majeure d’accidents de
la route. L’électroencéphalogr... more La somnolence des conducteurs est une cause majeure d’accidents de
la route. L’électroencéphalogramme (EEG) est considéré comme le prédicteur le plus robuste de cet état cérébral. Cet article propose une nouvelle méthode de détection de somnolence à l’aide d’une seule électrode, avec un potentiel d’implémentation temps réel. L’article présente d’abord une méthode originale pour déterminer le canal EEG le plus pertinent pour surveiller la somnolence, en utilisant l’analyse de covariance maximale. La seconde contribution consiste à développer une méthode d’apprentissage profond avec les signaux du canal déterminé. L’approche procède par extraction des caractéristiques spectrales du
signal. Ces caractéristiques sont utilisées avec un modèle de réseau récurrent à mémoire court et long terme (LSTM) pour détecter les états de somnolence. La méthode a été testée sur 12 sujets afin de discriminer les états de somnolence et d’alerte. Notre résultat principal est que le canal TP7, situé dans la région temporo-pariétale gauche, est le plus significatif. Cela correspond à une zone partagée entre la conscience spatiale et la navigation spatiale visuelle. Ce canal est aussi relié à la faculté de prudence. Malgré le faible volume de données, la méthode proposée permet de prédire la somnolence avec une précision de 75% et un délai moyen de 1.4 secondes. Ces résultats prometteurs mettent en lumière
des facteurs importants à considérer pour la surveillance de la somnolence.
Applied Sciences, 2022
the quality of the red raspberry accurately, automatically, and in real time. Raspberry trays wit... more the quality of the red raspberry accurately, automatically, and in real time. Raspberry trays with recently harvested fresh fruit enter the industry’s selection and quality control process to be categorized and subsequently their purchase price is determined. This selection is carried out from a sample of a complete batch to evaluate the quality of the raspberry. This database aims to solve one of the major problems in the industry: evaluating the largest amount of fruit possible and not a single sample. This major dataset enables researchers in various disciplines to develop practical machine-learning (ML) algorithms to improve red raspberry quality in the industry, by identifying different diseases and defects in the fruit, and by overcoming limitations by increasing the performance detection rate accuracy and reducing computation time. This database is made up of two packages and can be downloaded free from the Laboratory of Technological Research in Pattern Recognition repository at the Catholic University of the Maule. The RGB image package contains 286 raw original images with a resolution of 3948 × 2748 pixels from raspberry trays acquired during a typical process in the industry. Furthermore, the labeled images are available with the annotations for two diseases (86 albinism labels and 164 fungus rust labels) and two defects (115 over-ripeness labels, and 244 peduncle labels). The MATLAB code package contains three well-known ML methodological approaches, which can be used to classify and detect the quality of red raspberries. Two are statistical-based learning methods for feature extraction coupled with a conventional artificial neural network (ANN) as a classifier and detector. The first method uses four predictive learning from descriptive statistical measures, such as variance, standard deviation, mean, and median. The second method uses three predictive learning from a statistical model based on the generalized extreme value distribution parameters, such as location, scale, and shape. The third ML approach uses a convolution neural network based on a pre-trained fastest region approach (Faster R-CNN) that extracts its features directly from images to classify and detect fruit quality. The classification performance metric was assessed in terms of true and false positive rates, and accuracy. On average, for all types of raspberries studied, the following accuracies were achieved: Faster R-CNN 91.2%, descriptive statistics 81%, and generalized extreme value 84.5%. These performance metrics were compared to manual data annotations by industry quality control staff, accomplishing the parameters and standards of agribusiness. This work shows promising results, which can shed a new light on fruit quality standards methodologies in the industry.
Computers, 2020
Spike-and-wave discharge (SWD) pattern detection in electroencephalography (EEG) is a crucial sig... more Spike-and-wave discharge (SWD) pattern detection in electroencephalography (EEG) is a crucial signal processing problem in epilepsy applications. It is particularly important for overcoming time-consuming, difficult, and error-prone manual analysis of long-term EEG recordings. This paper presents a new method to detect SWD, with a low computational complexity making it easily trained with data from standard medical protocols. Precisely, EEG signals are divided into time segments for which the continuous Morlet 1-D wavelet decomposition is computed. The generalized Gaussian distribution (GGD) is fitted to the resulting coefficients and their variance and median are calculated. Next, a k-nearest neighbors (k-NN) classifier is trained to detect the spike-and-wave patterns, using the scale parameter of the GGD in addition to the variance and the median. Experiments were conducted using EEG signals from six human patients. Precisely, 106 spike-and-wave and 106 non-spike-and-wave signals were used for training, and 96 other segments for testing. The proposed SWD classification method achieved 95% sensitivity (True positive rate), 87% specificity (True Negative Rate), and 92% accuracy. These promising results set the path for new research to study the causes underlying the so-called absence epilepsy in long-term EEG recordings.
Journal of Biomedical Research, 2020
The two-point central difference is a common algorithm in biological signal processing and is par... more The two-point central difference is a common algorithm in biological signal processing and is particularly useful in analyzing physiological signals. In this paper, we develop a model-based classification method to detect epileptic seizures that relies on this algorithm to filter electroencephalogram (EEG) signals. The underlying idea was to design an EEG filter that enhances the waveform of epileptic signals. The filtered signal was fitted to a quadratic linear-parabolic model using the curve fitting technique. The model fitting was assessed using four statistical parameters, which were used as classification features with a random forest algorithm to discriminate seizure and non-seizure events. The proposed method was applied to 66 epochs from the Children Hospital Boston database. Results showed that the method achieved fast and accurate detection of epileptic seizures, with a 92% sensitivity, 96% specificity, and 94.1% accuracy.
Revista Argentina de Bioingeniería, 2019
Seizure detection plays a central role in most aspects of epilepsy care. Understanding the comple... more Seizure detection plays a central role in most aspects of epilepsy care. Understanding the complex epileptic signals system is a typical problem in electroencephalographic (EEG) signal processing. This problem requires different analysis to reveal the underlying behavior of EEG signals. An example of this is the non-linear dynamic: mathematical tools applied to biomedical problems with the purpose of extracting features or quantifying EEG data. In this work, we studied epileptic seizure detection independently in each brain rhythms from a multilevel 1D wavelet decomposition followed by the independent component analysis (ICA) representation of multivariate EEG signals. Next, the largest Lyapunov exponents (LLE) and their scaling given by its ± standard deviation are estimated in order to obtain the vectors to be used during the training and classification stage. With this information, a logistic regression classification is proposed with the aim of discriminating between seizure and non-seizure. Preliminary experiments with 99 epileptic events suggest that the proposed methodology is a powerful tool for detecting seizures in epileptic signals in terms of classification accuracy, sensitivity and specificity.
Digital Health Approach for Predictive, Preventive, Personalised and Participatory Medicine, 2019
The availability of electroencephalogram (EEG) data has opened up the possibility for new interes... more The availability of electroencephalogram (EEG) data has opened up the possibility for new interesting applications, such as epileptic seizure detection. The detection of epileptic activity is usually performed by an expert based on the analysis of the EEG data. This paper shows how a convolutional neural network (CNN) can be applied to EEG images for a full and accurate classification. The proposed methodology was applied on images reflecting the amplitude of the EEG data over all electrodes. Two groups are considered: (a) healthy subjects and (b) epileptic subjects. Classification results show that CNN has a potential in the classification of EEG signals, as well as the detection of epileptic seizures by reaching 99.48% of overall classification accuracy.
Digital Health Approach for Predictive, Preventive, Personalised and Participatory Medicine, 2019
This paper proposes a new algorithm for epileptic seizure onset detection in EEG signals. The alg... more This paper proposes a new algorithm for epileptic seizure onset detection in EEG signals. The algorithm relies on the measure of the entropy of observed data sequences. Precisely, the data is decomposed into different brain rhythms using wavelet multi-scale transformation. The resulting coefficients are represented using their generalized Gaussian distribution. The proposed algorithm estimates the parameters of the distribution and the associated entropy. Next, an ensemble bagging classifier is used to performs the seizure onset detection using the entropy of each brain rhythm, by discriminating between seizure and non-seizure. Preliminary experiments with 105 epileptic events suggest that the proposed methodology is a powerful tool for detecting seizures in epileptic signals in terms of classification accuracy, sensitivity and specificity.
Applied Medical Informatics, 2019
Epilepsy is an important public health issue. An appropriate epileptiform discharge pattern detec... more Epilepsy is an important public health issue. An appropriate epileptiform discharge pattern detection of this neurological disease is a typical problem in biomedical engineering. In this paper, a new method is proposed for spike-and-wave discharge pattern detection based on Kendall's Tau-b coefficient. The proposed approach is demonstrated on a real dataset containing spike-and-wave discharge signals, where our performance is evaluated in terms of high Specificity, rule in (SpPIn) with 94% for patient-specific spike-and-wave discharge detection and 83% for a general spike-and-wave discharge detection.
Neurología Argentina, 2018
To predict an epileptic event, means the ability to determine in advance the time of the seizure ... more To predict an epileptic event, means the ability to determine in advance the time of the seizure with the highest possible accuracy. A correct prediction benchmark for epilepsy events in clinical applications, is a typical problem in biomedical signal processing that help to an appropriate diagnosis and treatment of this disease. In this work we use Pearson's product-moment correlation coefficient from generalized Gaussian distribution parameters coupled with linear-based classifier to predict between seizure and non-seizure events in epileptic EEG signals. The performance in 36 epileptic events from 9 patients showing a good performance with 100% of effectiveness for sensitivity and specificity greater than 83% for seizures events in all brain rhythms. Pearson's test suggest that all brain rhythms are highly correlated in non-seizure events but no during the seizure events. This suggests that our model can be scaled with the Pearson product-moment correlation coefficient for the detection of epileptic seizures.
Biocybernetics and Biomedical Engineering, 2018
This paper presents a supervised classification method to accurately detect epileptic brain activ... more This paper presents a supervised classification method to accurately detect epileptic brain activity in real-time from electroencephalography (EEG) data. The proposed method has three main strengths: it has low computational cost, making it suitable for real-time implementation in EEG devices; it performs detection separately for each brain rhythm or EEG spectral band, following the current medical practices; and it can be trained with small datasets, which is key in clinical problems where there is limited annotated data available. This is in sharp contrast with modern approaches based on machine learning techniques, which achieve very high sensitivity and specificity but require large training sets with expert annotations that may not be available. The proposed method proceeds by first separating EEG signals into their five brain rhythms by using a wavelet filter bank. Each brain rhythm signal is then mapped to a low-dimensional manifold by using a generalized Gaussian statistical model; this dimensionality reduction step is computationally straightforward and greatly improves supervised classification performance in problems with little training data available. Finally, this is followed by parallel linear classifications on the statistical manifold to detect if the signals exhibit healthy or abnormal brain activity in each spectral band. The good performance of the proposed method is demonstrated with an application to paediatric neurology using 39 EEG recordings from the Children's Hospital Boston database, where it achieves an average sensitivity of 98%, specificity of 88%, and detection latency of 4 s, performing similarly to the best approaches from the literature.
Revista Argentina de Bioingeniería, 2018
It is estimated that 50% of all cardiovascular deaths are caused by a sudden cardiac arrest (SCA)... more It is estimated that 50% of all cardiovascular deaths are caused by a sudden cardiac arrest (SCA), which represents 15% of global mortality, and its main cause is ventricular fibrillation (VF). Therefore, it is of interest to design new methods capable to detect changes in heart rate (HR or RR interval) that could announce the beginning of an imminent fibrillation. In this work, an effective novel indicator, based on mean and standard deviation of Heart Rate Variability (HRV), was studied and used to develop an algorithm that predicts imminent VF with 100% sensitivity and 100% specificity. The study was based on 65 RR intervals signals. The algorithm's simplicity provides a quick-to-use implementation in a micro controller unit (MCU) for real-time VF detection, allowing its application in a variety of medical devices with electrocardiogram (ECG) modules.
Revista Argentina de Bioingeniería, 2018
Identify spike-and-waves patterns in epileptic signals is a typical problem in electroencephalogr... more Identify spike-and-waves patterns in epileptic signals is a typical problem in electroencephalographic (EEG) signal processing. In this paper we propose cross-correlation coupled with decision tree model as new method in order to assess and detect spike-and-wave discharges (SWD) in long-term epileptic signals. The proposed approach is demonstrated in terms of accuracy, sensitivity and specificity classification on real EEG signals using a database developed with medical annotations.
IFMBE Proceedings, Apr 7, 2017
This paper presents a statistical signal processing method for the characterization of EEG of pat... more This paper presents a statistical signal processing method for the characterization of EEG of patients suffering from epilepsy. A statistical model is proposed for the signals and the Kullback-Leibler divergence is used to study the differences between Seizure/Non-Seizure in patients suffering from epilepsy. Precisely, EEG signals are transformed into multivariate coefficients through multilevel 1D wavelet decomposition of different brain frequencies. The generalized Gaussian distribution (GGD) is shown to model precisely these coefficients. Patients are compared based on the analytical development of Kullback-Leibler divergence (KLD) of their corresponding GGD distributions. The method has been applied to a dataset of 18 epileptic signals of 9 patients. Results show a clear discrepancy between Seizure/Non-Seizure in epileptic signals, which helps in determining the onset of the seizure.
2016 IEEE 12th Image, Video, and Multidimensional Signal Processing Workshop (IVMSP)
The classification of epileptic seizure events in EEG signals is an important problem in biomedic... more The classification of epileptic seizure events in EEG signals is an important problem in biomedical engineering. In this paper we propose a Bayesian classification method for multivariate EEG signals. The method is based on a multilevel 2D wavelet decomposition that captures the distribution of energy across the different brain rhythms and regions, coupled with a generalised Gaussian statistical representation and a multivariate Bayesian classification scheme. The proposed approach is demonstrated on a challenging paediatric dataset containing both epileptic events and normal brain function signals, where it outperforms a state-of-the-art method both in terms of classification sensitivity and specificity.
Journal of Physics: Conference Series, 2016
Appropriate diagnosis and treatment of epilepsy is a main public health issue. Patients suffering... more Appropriate diagnosis and treatment of epilepsy is a main public health issue. Patients suffering from this disease often exhibit different physical characterizations, which result from the synchronous and excessive discharge of a group of neurons in the cerebral cortex. Extracting this information using EEG signals is an important problem in biomedical signal processing. In this work we propose a new algorithm for seizure onset detection and spread estimation in epilepsy patients. The algorithm is based on a multilevel 1-D wavelet decomposition that captures the physiological brain frequency signals coupled with a generalized gaussian model. Preliminary experiments with signals from 30 epilepsy crisis and 11 subjects, suggest that the proposed methodology is a powerful tool for detecting the onset of epilepsy seizures with his spread across the brain.
XIII Congreso Argentino de Acústica VII Jornadas de Acústica, Electroacústica y Áreas Vinculadas
A new digital timbre transformation that modifies harmonic sounds to inharmonic ones changing the... more A new digital timbre transformation that modifies harmonic sounds to inharmonic ones changing the frequency of the subharmonics is proposed. In order to perform the modifications, two digital signal processing techniques are discussed. The first one is a real time oriented algorithm and the second one is an analysis and resynthesis process. The results obtained using these techniques suggest that is possible to transform the timbre without having spectral leakage or significant loss in time resolution.
Fifth International Conference on Advances in New Technologies, Interactive Interfaces and Communicability ISBN:978.88.96.471.37.1, DOI: 10.978.8896471/371, Nov 11, 2014
Extracting information from scalp EEG signals is a challenging biomedical signal processing probl... more Extracting information from scalp EEG signals is a challenging biomedical signal processing problem that has a potentially strong impact in the diagnosis and treatment of numerous neurological conditions. In this work we study a new methodology for extracting information from EEG signals from patients suffering from epilepsy. The methodology is based on a multi- resolution wavelet representation and a statistical generalized Gaussian model, which provide a compact description of the time-frequency information in the EEG signal array. Preliminary experiments suggest that the information captured by the model is potentially useful for effectively detecting the onset of epileptic seizures, which is key for epilepsy diagnosis and treatment.
ECImag 4 Escuela y Workshop Argentino en Ciencias de las Imágenes
El presente trabajo hace un recorrido por diferentes formas de comunicar ciencia en forma de arte... more El presente trabajo hace un recorrido por diferentes formas de comunicar ciencia en forma de arte interactiva: la física con un juego no convencional de ping-pong que explora diferentes conceptos cinéticos, la música donde se expresa la gran diferencia entre el sonido estático y el sonido en movimiento, la lingüística donde se hace una analogía entre la geometría algebraica y las palabras y el movimiento corporal donde el usuario crea figuras y formas con las manos. Sus autores exploran diferentes tipos de interacción con el usuario y muestran posibilidades de presentar contenidos científicos en forma atractiva a un público general.
Uploads
Papers by Antonio Quintero-Rincón
la route. L’électroencéphalogramme (EEG) est considéré comme le prédicteur le plus robuste de cet état cérébral. Cet article propose une nouvelle méthode de détection de somnolence à l’aide d’une seule électrode, avec un potentiel d’implémentation temps réel. L’article présente d’abord une méthode originale pour déterminer le canal EEG le plus pertinent pour surveiller la somnolence, en utilisant l’analyse de covariance maximale. La seconde contribution consiste à développer une méthode d’apprentissage profond avec les signaux du canal déterminé. L’approche procède par extraction des caractéristiques spectrales du
signal. Ces caractéristiques sont utilisées avec un modèle de réseau récurrent à mémoire court et long terme (LSTM) pour détecter les états de somnolence. La méthode a été testée sur 12 sujets afin de discriminer les états de somnolence et d’alerte. Notre résultat principal est que le canal TP7, situé dans la région temporo-pariétale gauche, est le plus significatif. Cela correspond à une zone partagée entre la conscience spatiale et la navigation spatiale visuelle. Ce canal est aussi relié à la faculté de prudence. Malgré le faible volume de données, la méthode proposée permet de prédire la somnolence avec une précision de 75% et un délai moyen de 1.4 secondes. Ces résultats prometteurs mettent en lumière
des facteurs importants à considérer pour la surveillance de la somnolence.