Examinando por Autor "Santiago-Mozos, Ricardo"

Mostrando 1 - 4 de 4

An Automated Screening System for Tuberculosis
(IEEE, 2013-10) Santiago-Mozos, Ricardo; Pérez-Cruz, Fernando; Madden, Michael G.; Artés-Rodríguez, Antonio
Email Print Request Permissions Save to Project Automated screening systems are commonly used to detect some agent in a sample and take a global decision about the subject (e.g. ill/healthy) based on these detections. We propose a Bayesian methodology for taking decisions in (sequential) screening systems that considers the false alarm rate of the detector. Our approach assesses the quality of its decisions and provides lower bounds on the achievable performance of the screening system from the training data. In addition, we develop a complete screening system for sputum smears in tuberculosis diagnosis, and show, using a real-world database, the advantages of the proposed framework when compared to the commonly used count detections and threshold approach.
Estimation of Reference Indices of Left Ventricular Chamber Function from Echocardiographic Images with Multidimensional Kernel Methods
(IEEE, 2012-09-09) Santiago-Mozos, Ricardo; Rojo-Álvarez, José Luis; Antoranz, J Carlos; Rodriguez, Daniel; Desco, Mar; Barrio, Alicia; Benito, Yolanda; Yotti, Raquel; Bermejo, Javier
Advanced nonlinear estimation methods can compete with their linear counterparts for the estimation of left ventricular (LV) function indices from color-Doppler M-mode images. We benchmarked three methods: Support Vector Regression, Partial Least Squares and Principal Component Regression using linear and non-linear (Gaussian) kernels. Two reference indices were directly estimated from the images, namely, the peak-systolic elastance (Emax) and the time-constant of LV relaxation (Tau). We found linear methods performing slightly better for predicting Emax, an easier task, but they were outperformed by non-linear procedures when predicting Tau, a harder estimation problem.
On feature extraction for noninvasive kernel estimation of left ventricular chamber function indices from echocardiographic images
(2015-01-30) Santiago-Mozos, Ricardo; Rojo-Álvarez, José Luis; Antoranz, J. Carlos; Rodríguez-Pérez, Daniel; Yotti, Raquel; Bermejo, Javier; Elsevier
Two reference indices used to characterize left ventricular (LV) global chamber function are end-systolic peak elastance (EmaxEmax) and the time-constant of relaxation rate (¿ ). However, these two indices are very difficult to obtain in the clinical setting as they require invasive high-fidelity catheterization procedures. We have previously demonstrated that it is possible to approximate these indices noninvasively by digital processing color-Doppler M-mode (CDMM) images. The aim of the present study was twofold: (1) to study which feature extraction from linearly reduced input spaces yields the most useful information for the prediction of the haemodynamic variables from CDMM images; (2) to verify whether the use of nonlinear versions of those linear methods actually improves the estimation. We studied the performance and interpretation of different linearly transformed input spaces (raw image, discrete cosine transform (DCT) coefficients, partial least squares, and principal components regression), and we compared whether nonlinear versions of the above methods provided significant improvement in the estimation quality. Our results showed that very few input features suffice for providing a good (medium) quality estimator for EmaxEmax (for ¿), which can be readily interpreted in terms of the measured flows. Additional covariates should be included to improve the prediction accuracy of both reference indices, but especially in ¿ models. The use of efficient nonlinear kernel algorithms does improve the estimation quality of LV indices from CDMM images when using DCT input spaces that capture almost all energy.
On the Early Detection of Perinatal Hypoxia with Information-Theory based Methods
(2013-09-26) Santiago-Mozos, Ricardo; García-Vizuete, Beatriz; Lillo-Castellano, José María; Rojo-Álvarez, José Luis; Martín-Caballero, Carlos
Perinatal hypoxia is a severe condition that may harm fetus organs permanently. When the fetus brain is partially deprived from oxygen, the control of the fetal heart rate (FHR) is affected. We hypothesized that advanced processing of the FHR can reveal whether the fetus is under perinatal hypoxia. We analyzed FHR morphology with normalized compression distance (NCD) that compares two arbitrary sequences and outputs their dissimilarity. This parameter-free measure exploits linear and non-linear relations in the data and allows the comparison of sequences of different sizes. It was applied to raw FHR sequences and to a set of statistics computed from them (e.g. moments on 5 minutes signal windows). We classified the cases from the NCD dissimilarity matrix by using a simple nearest neighbor classifier and leave-one-out cross-validation. Best results in a database with 26 FHR recordings (13 controls and 13 cases) were provided by the central moment of order 3 calculated over sliding windows of 5 minutes on the interval from 4 to 3 hours to delivery. The resulting accuracy was 0.88 with sensitivity 0.92 and specificity 0.85.