Abstract
Novel data collection tools, methods and new techniques in biotechnology can facilitate improved health strategies that are customised to each individual. One key challenge to achieve this is to take advantage of the massive volumes of personal anonymous data, relating each profile to health and disease, while accounting for high diversity in individuals, populations and environments. These data must be analysed in unison to achieve statistical power, but presently cohort data repositories are scattered, hard to search and integrate, and data protection and governance rules discourage central pooling.

Abstract

In the last decade the advances in wearable technology have driven and transformed performance monitoring in fitness and wellness applications, surveillance in extreme (working) conditions, and management of chronic diseases. These innovations have opened a whole new perspective on health and social care, challenged by vast expenditures in ageing societies.

Abstract
The Sternum is a human bone located in the anterior area of the thoracic cage. It is present in most of the axial cuts provided from the Magnetic Resonance Imaging (MRI) acquisitions. used in the medical field. Detecting the Sternum is relevant as it contains rigid key-points for 3D model reconstructions, assisting in the planning and evaluation of several surgical procedures, and for atlas development by segmenting structures in anatomical proximity. In the absence of applicable approaches for this specific problem. this paper focuses on two distinct automated methods for Sternum segmentation in MRI. The first. relies on K-Means (Clustering) to perform the segmentation, while the second encompasses the closed Minimum Path over the elliptical transformation of Gradient images. A dataset of 14 annotated acquisitions was used for evaluation. The results favored the Gradient approach over Clustering.

Abstract
Visual exploration of the larynx represents a relevant technique for the early diagnosis of laryngeal disorders. However, visualizing an endoscopy for finding abnormalities is a time-consuming process, and for this reason much research has been dedicated to the automatic analysis of endoscopic video data. In this work we address the particular task of discriminating among informative laryngoscopic frames and those that carry insufficient diagnostic information. In the latter case, the goal is also to determine the reason for this lack of information. To this end, we analyze the possibility of training three different state-of-the-art Convolutional Neural Networks, but initializing their weights from configurations that have been previously optimized for solving natural image classification problems. Our findings show that the simplest of these three architectures not only is the most accurate (outperforming previously proposed techniques), but also the fastest and most efficient, with the lowest inference time and minimal memory requirements, enabling real-time application and deployment in portable devices.

Abstract
The segmentation of blood vessels in medical images has been heavily studied, given its impact in several clinical practices. Deep Learning methods have been applied to supervised segmentation of blood vessels, mainly the retinal ones due to the availability of manual annotations. Despite their success, they typically minimize the Binary Cross Entropy loss, which does not penalize topological mistakes. These errors are relevant in graph-like structures such as blood vessel trees, as a missing segment or an inadequate merging or splitting of branches, may severely change the topology of the network and put at risk the extraction of vessel pathways and their characterization. In this paper, we propose an end-to-end network design comprising a cascade of a typical segmentation network and a Variational Auto-Encoder which, by learning a rich but compact latent space, is able to correct many topological incoherences. Our experiments in three of the most commonly used retinal databases, DRIVE, STARE, and CHASEDB1, show that the proposed model effectively learns representations inducing better segmentations in terms of topology, without hurting the usual pixel-wise metrics.

Abstract
Abstract Background Seizures commonly affect the heart rate and its variability. The increased interest in this area of research is related to the possible connection with sudden unexpected death in epilepsy (SUDEP). Generalized tonic-clonic seizures (GTCS) are reported as the most consistent risk factor for SUDEP. However, the general risk of seizures (and their type) on cardiac function still remains uncertain. Purpose To evaluate the influence of seizure type (GTCS vs non-GTCS) on ictal and early post-ictal Heart Rate Variability (HRV) in patients with refractory epilepsy. Methods From January 2015 to July 2018, we prospectively evaluated 121 patients admitted to our institution's Epilepsy Monitoring Unit with focal resistant epilepsy. All patients underwent a 48-hour Holter recording. We included only patients who had both GTCS and non-GTCS during the recording and selected the first seizure of each type to analyze. HRV (AVNN, SDNN, RMSSD, pNN50, and LF/HF) was evaluated by analyzing 5-min-ECG epochs, starting with the seizure onset (ictal and early post-ictal period). The study was approved by our Institution Ethics Committee and all patients gave informed consent. Results Fourteen patients were included (7 Females, 4 patients with Temporal Lobe Epilepsy). The median age was 39 years (min-max, 18–57). Thirty-six percent presented cardiovascular risk factors without known cardiac disease. A significant statistical reduction was found for AVNN (p=0.013), RMSSD (p=0.008), pNN50 (p=0.005) and HF (p=0.003), during GTCS when compared with non-GTCS (Wilcoxon test, p<0.05; two tailed). Conclusion Our study shows a significant reduced vagal tone during GTCS when compared with non-GTCS. Hence, GTCS had a more pronounced impact on HRV changes than other seizure types, which can be associated with higher SUDEP risk after GTCS.