RESUMO
Chest radiographs allow for the meticulous examination of a patient's chest but demands specialized training for proper interpretation. Automated analysis of medical imaging has become increasingly accessible with the advent of machine learning (ML) algorithms. Large labeled datasets are key elements for training and validation of these ML solutions. In this paper we describe the Brazilian labeled chest x-ray dataset, BRAX: an automatically labeled dataset designed to assist researchers in the validation of ML models. The dataset contains 24,959 chest radiography studies from patients presenting to a large general Brazilian hospital. A total of 40,967 images are available in the BRAX dataset. All images have been verified by trained radiologists and de-identified to protect patient privacy. Fourteen labels were derived from free-text radiology reports written in Brazilian Portuguese using Natural Language Processing.
Assuntos
Algoritmos , Processamento de Linguagem Natural , Radiografia Torácica , Brasil , Humanos , Raios XRESUMO
The Stark effect is one of the most efficient mechanisms to manipulate many-body states in nanostructured systems. In mono- and few-layer transition metal dichalcogenides, it has been successfully induced by optical and electric field means. Here, we tune the optical emission energies and dissociate excitonic states in MoSe2 monolayers employing the 220 MHz in-plane piezoelectric field carried by surface acoustic waves. We transfer the monolayers to high dielectric constant piezoelectric substrates, where the neutral exciton binding energy is reduced, allowing us to efficiently quench (above 90%) and red-shift the excitonic optical emissions. A model for the acoustically induced Stark effect yields neutral exciton and trion in-plane polarizabilities of 530 and 630 × 10-5 meV/(kV/cm)2, respectively, which are considerably larger than those reported for monolayers encapsulated in hexagonal boron nitride. Large in-plane polarizabilities are an attractive ingredient to manipulate and modulate multiexciton interactions in two-dimensional semiconductor nanostructures for optoelectronic applications.
RESUMO
Continuous flow enteral fluid therapy with isotonic and hypotonic enteral electrolyte solutions are as safe and effective as intravenous fluid therapy. The aim of this study was to carry out a comparative assessment between continuous flow enteral and intravenous (IV) fluid therapy in adult experimentally dehydrated horses. Six experimentally dehydrated adult mares were used in a study carried out in a 6 × 3 crossover design, which each animal received three different treatments (isotonic enteral fluid therapy-EsISO, hypotonic enteral fluid therapy-EsHYPO and intravenous fluid therapy with Lactate Ringer Solution-LR IV, all in continuous flow). Solutions were administered at a rate of 15 mL-1.kg-1.h-1 for 8 h, after 36 h of water and food deprivation. Serum and urinary biochemical assessment; urinary volume, pH and specific gravity; and blood gas analysis were measured at -36, 0, 2, 4, 6, and 8 h. The dehydration period (DP) caused discrete hydroelectrolytic and acid base imbalances. The EsISO, EsHYPO and LR IV increased blood volume. Enteral solutions restored the imbalances yielded by the DP and all treatments increased urine volume. Also, the EsHYPO and LR IV showed no effects in acid base balance, while EsISO showed slightly acidifying effect. The present study certifies the efficacy and safety of isotonic and hypotonic continuous flow enteral fluid therapy in comparison to IV fluid therapy in dehydrated horses.