arXiv subjects

Pedro Costa

Publications and source records attributed to Pedro Costa.

At least 55 records · Page 3Linked to original sources

arXiv2020

Improving Lesion Segmentation for Diabetic Retinopathy using Adversarial Learning

Diabetic Retinopathy (DR) is a leading cause of blindness in working age adults. DR lesions can be challenging to identify in fundus images, and automatic DR detection systems can offer strong clinical value. Of the publicly available labeled datasets for DR, the Indian Diabetic Retinopathy Image Dataset (IDRiD) presents retinal fundus images with pixel-level annotations of four distinct lesions: microaneurysms, hemorrhages, soft exudates and hard exudates. We utilize the HEDNet edge detector to solve a semantic segmentation task on this dataset, and then propose an end-to-end system for pixel-level segmentation of DR lesions by incorporating HEDNet into a Conditional Generative Adversarial Network (cGAN). We design a loss function that adds adversarial loss to segmentation loss. Our experiments show that the addition of the adversarial loss improves the lesion segmentation performance over the baseline.

eess.IV↗

arXiv2020

Learned Pre-Processing for Automatic Diabetic Retinopathy Detection on Eye Fundus Images

Diabetic Retinopathy is the leading cause of blindness in the working-age population of the world. The main aim of this paper is to improve the accuracy of Diabetic Retinopathy detection by implementing a shadow removal and color correction step as a preprocessing stage from eye fundus images. For this, we rely on recent findings indicating that application of image dehazing on the inverted intensity domain amounts to illumination compensation. Inspired by this work, we propose a Shadow Removal Layer that allows us to learn the pre-processing function for a particular task. We show that learning the pre-processing function improves the performance of the network on the Diabetic Retinopathy detection task.

eess.IV↗

arXiv2020

Magnetic field dependent 't Hooft determinant extended Nambu--Jona-Lasinio Model

We study the implications of recent lattice QCD results for the magnetic field dependence of the quarks dynamical masses on the 't Hooft determinant extended Nambu--Jona-Lasinio Model in the light and strange quark sectors (\emph{up}, \emph{down} and \emph{strange}). The parameter space is constrained at vanishing magnetic field, using the quarks dynamical masses and the meson spectra, whereas at non-vanishing magnetic field strength the dependence of the dynamical masses of two of the quark flavors is used to fit a magnetic field dependence on the model couplings, both the four-fermion Nambu--Jona-Lasinio interaction and the six-fermion 't Hooft flavor determinant. We found that this procedure reproduces the inverse magnetic catalysis, and the strength of the scalar coupling decreases with the magnetic field, while the strength of the six-fermion 't Hooft flavor determinant increases with the magnetic field.

hep-ph↗

arXiv2020

Quark and Polyakov-loop correlations in effective models at zero and non-vanishing density

The aim of this work is to shed light on some lesser known aspects of Polyakov-loop--extended chiral models (namely the Polyakov-loop extended Nambu--Jona-Lasinio and Quark-Meson models), especially on the correlation of the quark sector with the Polyakov loop. We show that the ordering of chiral and Polyakov-loop transitions and their difference in temperature as seen in lattice QCD calculations could be realized with a critical scale of the Polyakov-loop potential that is larger than the one in pure gauge theory. The comparison of the results for the Polyakov-loop susceptibility obtained using the self-consistent medium-dependent quark mass with those obtained while keeping these masses at a fixed value allows to disentangle chiral-symmetry restoration and center-symmetry breaking effects. Furthermore, a confined chirally restored phase is identified by a plateau in the quark contribution to thermodynamics and by sigma and pion spectral functions that coincide but have a small width. We also discuss that, for some large chemical potential values, the explicit center-symmetry breaking is so strong that statistical deconfinement is realized at infinitely small temperatures. Both the missing sensitivity of the Polyakov loop to the quark mass, except at close to the chiral transition, and the Polyakov loop being zero at zero temperature at all chemical potentials, can be interpreted as indications of a missing mechanism which accounts for the quark back-reaction on the Polyakov loop.

hep-ph↗

arXiv2020

One-meson-loop NJL model: Effect of collective and noncollective excitations on the quark condensate at finite temperature

We explore the effect of including quantum fluctuations in the two flavor Nambu$-$Jona-Lasinio model at finite temperature. This is accomplished, in a symmetry preserving way, by including collective and noncollective modes in the one-meson-loop gap equation which originate from poles and branch cuts in the complex plane, respectively. The inclusion of a boson cutoff, $Λ_b$, is necessary to regularize the meson-loop momenta. This new parameter is used to study the influence of going beyond the usual mean field approximation in the quark condensate. As the temperature increases, chiral symmetry tends to get restored, the collective modes melt and only noncollective modes contribute to the quark condensate. With the inclusion of such modes, the quark condensate at finite temperature has a different behavior than at mean field level that will be explored.

hep-ph↗

arXiv2020

A volume-of-fluid method for interface-resolved simulations of phase-changing two-fluid flows

We present a numerical method for interface-resolved simulations of evaporating two-fluid flows based on the volume-of-fluid (VoF) method. The method has been implemented in an efficient FFT-based two-fluid Navier-Stokes solver, using an algebraic VoF method for the interface representation, and extended with the transport equations of thermal energy and vaporized liquid mass for the single-component evaporating liquid in an inert gas. The conservation of vaporizing liquid and computation of the interfacial mass flux are performed with the aid of a reconstructed signed-distance field, which enables the use of well-established methods for phase change solvers based on level-set methods. The interface velocity is computed with a novel approach that ensures accurate mass conservation, by constructing a divergence-free extension of the liquid velocity field onto the entire domain. The resulting approach does not depend on the type of interface reconstruction (i.e.\ can be employed in both algebraic and geometrical VoF methods). We extensively verified and validated the overall method against several benchmark cases, and demonstrated its excellent mass conservation and good overall performance for simulating evaporating two-fluid flows in two and three dimensions.

physics.comp-ph↗

arXiv2019

A FFT-based finite-difference solver for massively-parallel direct numerical simulations of turbulent flows

We present an efficient solver for massively-parallel direct numerical simulations of incompressible turbulent flows. The method uses a second-order, finite-volume pressure-correction scheme, where the pressure Poisson equation is solved with the method of eigenfunction expansions. This approach allows for very efficient FFT-based solvers in problems with different combinations of homogeneous pressure boundary conditions. Our algorithm explores all combinations of pressure boundary conditions valid for such a solver, in a single, general framework. The method is implemented in a 2D pencil-like domain decomposition, which enables efficient massively-parallel simulations. The implementation was validated against different canonical flows, and its computational performance was examined. Excellent strong scaling performance up to $10^4$ cores is demonstrated for a domain with $10^9$ spatial degrees of freedom, corresponding to a very small wall-clock time/time step. The resulting tool, CaNS, has been made freely available and open-source.

physics.flu-dyn↗

arXiv2019

Aspects of isentropic trajectories in chiral effective models

The evolution of the fireball in heavy ion collisions is an isentropic process, meaning that it follows a trajectory of constant entropy per baryon in the phase diagram of the strong interaction. Responsible for the collective acceleration of the fireball is the speed of sound of the system, while fluctuations of conserved charges are encoded in quark-number susceptibilities: together, they leave their imprint in final observables. Here, this isentropic evolution will be analysed within chiral effective models that account for both chiral and center symmetry breaking, two central aspects of QCD. Our discussion focusses on the impact on the isentropic trajectories of the treatment of high-momentum modes, of the meson contribution to thermodynamics and of the number of quark flavours.

hep-ph↗

arXiv2019

Phase Diagram, Scalar-Pseudoscalar Meson Behavior and Restoration of Symmetries in (2+1) Polyakov-Nambu-Jona-Lasinio Model

We explore the phase diagram and the modification of mesonic observables in a hot and dense medium using the (2+1) Polyakov-Nambu-Jona-Lasinio model. We present the phase diagram in the ($T,\,μ_B$)-plane, with its isentropic trajectories, paying special attention to the chiral critical end point (CEP). Chiral and deconfinement transitions are examined. The modifications of mesonic observables in the medium are explored as a tool to analyze the effective restoration of chiral symmetry for different regions of the phase diagram. It is shown that the meson masses, namely that of the kaons, change abruptly near the CEP, which can be relevant for its experimental search.

hep-ph↗

arXiv2018

Magnetized QCD phase diagram: net-baryon susceptibilities

Employing the Polyakov extended Nambu-Jona-Lasinio model, we determine the net-baryon number fluctuations of magnetized three-flavor quark matter. We show that the magnetic field changes the nature of the strange quark transition from crossover to first-order at low temperatures. In fact, the strange quark undergoes multiple first-order phase transitions and several critical end points emerge in the phase diagram.

hep-ph↗

arXiv2018

UOLO - automatic object detection and segmentation in biomedical images

We propose UOLO, a novel framework for the simultaneous detection and segmentation of structures of interest in medical images. UOLO consists of an object segmentation module which intermediate abstract representations are processed and used as input for object detection. The resulting system is optimized simultaneously for detecting a class of objects and segmenting an optionally different class of structures. UOLO is trained on a set of bounding boxes enclosing the objects to detect, as well as pixel-wise segmentation information, when available. A new loss function is devised, taking into account whether a reference segmentation is accessible for each training image, in order to suitably backpropagate the error. We validate UOLO on the task of simultaneous optic disc (OD) detection, fovea detection, and OD segmentation from retinal images, achieving state-of-the-art performance on public datasets.

cs.CV↗

arXiv2018

MedAL: Deep Active Learning Sampling Method for Medical Image Analysis

Deep learning models have been successfully used in medical image analysis problems but they require a large amount of labeled images to obtain good performance.Deep learning models have been successfully used in medical image analysis problems but they require a large amount of labeled images to obtain good performance. However, such large labeled datasets are costly to acquire. Active learning techniques can be used to minimize the number of required training labels while maximizing the model's performance.In this work, we propose a novel sampling method that queries the unlabeled examples that maximize the average distance to all training set examples in a learned feature space. We then extend our sampling method to define a better initial training set, without the need for a trained model, by using ORB feature descriptors. We validate MedAL on 3 medical image datasets and show that our method is robust to different dataset properties. MedAL is also efficient, achieving 80% accuracy on the task of Diabetic Retinopathy detection using only 425 labeled images, corresponding to a 32% reduction in the number of required labeled examples compared to the standard uncertainty sampling technique, and a 40% reduction compared to random sampling.

cs.CV↗

arXiv2018

Experimental investigation of turbulent suspensions of spherical particles in a square duct

We report experimental observations of turbulent flow with spherical particles in a square duct. Three particle sizes namely: $2H/d_{p}$ = 40, 16 and 9 ($2H$ being the duct full height and $d_{p}$ being the particle diameter) are investigated. The particles are nearly neutrally-buoyant with respect to the suspending fluid. Refractive Index Matched - Particle Image Velocimetry is used for fluid velocity measurement even at the highest particle volume fraction (20%) and Particle Tracking Velocimetry for the particle velocity statistics for the flows seeded with particles of the two largest sizes, whereas only pressure measurements are reported for the smallest particles. Settling effects are seen at the lowest bulk Reynolds number $Re_{2H}\approx$ 10000 whereas, at the highest $Re_{2H}\approx$ 27000, particles are in almost full suspension. The friction factor of the suspensions is found to be significantly larger than that of single-phase duct flow at the lower $Re_{2H}$ investigated; however, the difference decreases when increasing the flow rate and the total drag approaches the values of the single phase flow at the higher Reynolds number considered. The pressure drop is found to decrease with the particle diameter for volume fractions lower than $ϕ$ = 10% for nearly all $Re_{2H}$. However, at the highest volume fraction $ϕ$ = 20%, we report a peculiar non-monotonic behavior: the pressure drop first decreases and then increases with increasing particle size. The decrease of the turbulent drag with particle size at the lowest volume fractions is related to an attenuation of the turbulence. The drag increase for the two largest particle sizes at $ϕ$ = 20%, however, occurs despite this large reduction of the turbulent stresses, and it is therefore due to significant particle-induced stresses.

physics.flu-dyn↗

arXiv2018

Net baryon-number fluctuations in magnetized quark matter

The kurtosis and skewness of net baryon-number fluctuations are studied for the magnetized phase diagram of three-flavor quark matter within the Polyakov extended Nambu$-$Jona-Lasinio model. Two models with magnetic catalysis and inverse magnetic catalysis are considered. Special attention is given to their behavior in the neighborhood of the light and strange critical end points (CEPs). Several isentropic trajectories that come close the CEPs are studied in order to analyze possible signatures of a CEP in the presence of external magnetic fields. The effect of the magnetic field on the velocity of sound, $v_s^2$, when both the light and strange CEPs are approached from the crossover region is also investigated by calculating their temperature and baryon chemical potential dependencies at fixed distances from these CEPs. Regions with large fluctuations but no CEP in nonmagnetized matter develop a CEP under the action of a strong magnetic field. Besides, the Landau quantization of the quark trajectories may result in the appearance of extra CEPs, in particular, in the strange sector for strong magnetic fields, identifiable by the net baryon-number fluctuations. Stiffer (smoother) fluctuations in the region of the CEP are characteristic of models that do not predict (do predict) the inverse magnetic catalysis at zero chemical potential. Particularly interesting is the ratio $χ^4_B/χ^2_B$ that has a more pronounced peak structure, indicating that it is eventually a more convenient probe for the search of a CEP. The speed of sound shows a much richer structure in magnetized quark matter and allows one to identify both chiral and deconfinement transitions.

hep-ph↗

arXiv2018

The presence of a critical end point in the QCD phase diagram from the net-baryon number fluctuations

The net-baryon number fluctuations for three-flavor quark matter are computed within the Polyakov extended Nambu$-$Jona-Lasinio model. Two models with vanishing and nonvanishing vector interactions are considered. While the former predicts a critical end point (CEP) in the phase diagram, the latter predicts no CEP. We show that the nonmonotonic behavior of the susceptibilities in the phase diagram is still present even in the absence of a CEP. Therefore, from the nonmonotonic behavior of the susceptibilities solely, one cannot assume the existence of a CEP. We analyze other possible properties that may distinguish the two scenarios, and determine the behavior of the net-baryon number fluctuations and the velocity of sound along several isentropes, with moderate and small values. It is shown that the value of the susceptibilities ratios and the velocity of sound at two or three isentropic lines could possibly allow to distinguish both scenarios, a phase diagram with or without CEP. Smoother behaviors of these quantities may indicate the nonexistence of a CEP. We also discuss the critical behavior of the strange sector.

hep-ph↗

arXiv2018

Effects of the finite particle size in turbulent wall-bounded flows of dense suspensions

We use interface-resolved simulations to study finite-size effects in turbulent channel flow of neutrally-buoyant spheres. Two cases with particle sizes differing by a factor of 2, at the same solid volume fraction of 20% and bulk Reynolds number are considered. These are complemented with two reference single-phase flows: the unladen case, and the flow of a Newtonian fluid with the effective suspension viscosity of the same mixture in the laminar regime. As recently highlighted in Costa et al. (PRL 117, 134501), a particle-wall layer is responsible for deviations of the statistics from what is observed in the continuum limit where the suspension is modeled as a Newtonian fluid with an effective viscosity. Here we investigate the fluid and particle dynamics in this layer and in the bulk. In the particle-wall layer, the near wall inhomogeneity has an influence on the suspension micro-structure over a distance proportional to the particle size. In this layer, particles have a significant (apparent) slip velocity that is reflected in the distribution of wall shear stresses. This is characterized by extreme events (both much higher and much lower than the mean). Based on these observations we provide a scaling for the particle-to-fluid apparent slip velocity as a function of the flow parameters. We also extend the flow scaling laws in to second-order Eulerian statistics in the homogeneous suspension region away from the wall. Finite-size effects in the bulk of the channel become important for larger particles, while negligible for lower-order statistics and smaller particles. Finally, we study the particle dynamics along the wall-normal direction. Our results suggest that 1-point dispersion is dominated by particle-turbulence (and not particle-particle) interactions, while differences in 2-point dispersion and collisional dynamics are consistent with a picture of shear-driven interactions.

physics.flu-dyn↗

arXiv2018

Multiple critical end points in magnetized three flavor quark matter

The magnetized phase diagram for three-flavor quark matter is studied within the Polyakov extended Nambu--Jona-Lasinio model. The order parameters are analyzed with special emphasis on the strange quark condensate. We show that the presence of an external magnetic field induces several critical endpoints (CEPs) in the strange sector, which arise due to the multiple phase transitions that the strange quark undergoes. The spinodal and binodal regions of the phase transitions are shown to increase with external magnetic field strength. The influence of strong magnetic fields on the isentropic trajectories around the several CEPs is analyzed. A focusing effect is observed on the region towards the CEPs that are related with the strange quark phase transitions. Compared to the chiral transitions, the deconfinement transition turns out to be less sensitive to the external magnetic field and the crossover nature is preserved over the whole phase diagram.

hep-ph↗

arXiv2017

Magnetized QCD phase diagram: critical end points for the strange quark phase transition driven by external magnetic fields

In this work we examine possible effects of an external magnetic field in the strongly interacting matter phase diagram. The study is performed using the Polyakov-Nambu-Jona-Lasinio model. Possible consequences of the inverse magnetic catalysis effect on the phase diagram at both finite chemical potential and temperature are analyzed. We devote special emphasis on how the location of the multiple critical end points (CEPs) change in a magnetized medium: the presence of an external magnetic field induces several CEPs in the strange sector, which arise due to the multiple phase transitions that the strange quark undergoes. We also study the deconfinement transition which turns out to be less sensitive to the external magnetic field when compared to the quark phase transitions. The crossover nature of the deconfinement is preserved over the whole phase diagram.

hep-ph↗