Abstract
A method to detect juxta-pleural nodules with radius smaller than 5mm is presented. The intensity difference between nodules and parenchymal tissue as well as the nodules' natural roundness are exploited. Solid nodules are detected by selecting an appropriate threshold over a sliding window, whereas sub-solid/non-solid nodules are enhanced using multi-scale Laplacian-of-Gaussian filters. The 2D-wise outputs are combined to 3D, producing a final candidate list. False positive reduction is achieved with fixed rules and supervised learning. The achieved sensitivity is 57% with 4 false positives/scan, increasing to 62% if only solid nodules are considered. © 2017 SPIE.

Abstract
The ability to detect pathological changes early and in a non-invasive way represents important advantages in the medical field. Diagnosis should become less intrusive, more accurate and less expensive in order to implement in the clinical routine. Infrared thermography has the advantages of being non-invasive, fast, reliable, capable of producing multiple recordings in short intervals, and absolutely safe for patients and clinicians. Thermographic image (TI) came to be an extensively studied technique to quantify sensitive changes in skin temperature in relation to certain diseases: early in the pathological process (lesions, inflammation and infection) the circulation fluxes are altered and, consequently, the tissues’ temperature is reflected in thermography pattern, before structural or functional changes can be observed. This technique proved to be able to give relevant clinical information, such as breast cancer, foot disease in diabetes, rheumatoid arthritis and sports injuries. Monitoring the temperature profile of a patient will allow understanding the physiological evolution of some diseases or monitoring the pharmacologic therapy effect. However, the high cost of this technology and the small number of commercial solutions do not allow a general implementation in the clinical environmental. The future direction is the combination of this technique with the other images techniques in order to add clinical information for a more reliable diagnostic.

Abstract
Background: Deep Brain Stimulation (DBS) of the Globus pallidus internus (GPi) is gold standard treatment in medically refractory dystonia. Recent evidence indicates that stimulation effects are also due to axonal modulation and affection of a fibre network. For the GPi, the pallidothalamic tracts are known to be the major motor efferent pathways. The aim of this study is to explore the anatomic vicinity of these tracts and DBS electrodes in dystonia applying diffusion tractography. Methods: Diffusion MRI was acquired in ten patients presenting for DBS for dystonia. We applied both a conventionally used probabilistic tractography algorithm (FSL) as well as a probabilistic streamline tracking approach, based on constrained spherical deconvolution and particle filtering with anatomic priors, to the datasets. DBS electrodes were coregistered to the diffusion datasets. Results: We were able to delineate the pallidothalamic tracts in all patients. Using the streamline approach, we were able to distinguish between the two sub-components of the tracts, the ansa lenticularis and the fasciculus lenticularis. Clinically efficient DBS electrodes displayed a close anatomic vicinity pathway of the pallidothalamic tracts, and their course was consistent with previous tracer labelling studies. Although we present only anatomic data, we interpret these findings as evidence of the possible involvement of fibre tracts to the clinical effect in DBS. Electro-physiological intraoperative recordings would be needed to complement our findings. In the future, a clear and individual delineation of the pallidothalamic tracts could optimize the stereotactic process of optimal electrode localization. (C) 2016 Wiley Periodicals, Inc.

Abstract

Abstract

Abstract
Assessment of retinal vessels is fundamental for the diagnosis of many disorders such as heart diseases, diabetes and hypertension. The imaging of retina using advanced fundus camera has become a standard in computer-assisted diagnosis of opthalmic disorders. Modern cameras produce high quality color digital images, but during the acquisition process the light reflected by the retinal surface generates a luminosity and contrast variation. Irregular illumination can introduce severe distortions in the resulting images, decreasing the visibility of anatomical structures and consequently demoting the performance of the automated segmentation of these structures. In this paper, a novel approach for illumination correction of color fundus images is proposed and applied as preprocessing step for retinal vessel segmentation. Our method builds on the connection between two different phenomena, shadows and haze, and works by removing the haze from the image in the inverted intensity domain. This is shown to be equivalent to correct the nonuniform illumination in the original intensity domain. We tested the proposed method as preprocessing stage of two vessel segmentation methods, one unsupervised based on mathematical morphology, and one supervised based on deep learning Convolutional Neural Networks (CNN). Experiments were performed on the publicly available retinal image database DRIVE. Statistically significantly better vessel segmentation performance was achieved in both test cases when illumination correction was applied.