Abstract

Abstract
Glaucoma is a silent disease that shows symptoms when severe, leading to partial vision loss or irreversible blindness. Early screening permits treating patients in time. For glaucoma screening, retinal images are very important since they enable the observation of initial glaucoma lesions, which typically begins with the cupping formation in the optic disc (OD). In clinical settings, practical indicators such as Cup-to-Disc Ratio (CDR) are frequently used to evaluate the presence and stage of glaucoma. The ratio between the cup and the optic disc can be measured using the vertical or horizontal diameter, or the area of the two. Mass screening programs are limited by the high costs of specialised teams and equipment. Current deep learning (DL) methods can assist the glaucoma mass screening, lower the cost and allow it to be extended to larger populations. With DL methods in the OD and optic cup (OC) segmentation, is possible to evaluate the presence of glaucoma in the patient more quickly based on cupping formation in the OD, using CDR. In this work, is assessed the contribution of Multi-Class and Single-Class segmentation methods for glaucoma screening using the 3 types of CDR. U-Net architecture is trained using transfer learning models (Inception V3 and Inception ResNet V2) to segment the OD and OC and then evaluate glaucoma prediction based on different types of CDRs indicators. The models were trained and evaluated on main public known databases (REFUGE, RIM-ONE r3 and DRISHTI-GS). The segmentation of both OD and OC reach Dice over 0.8 and IoU above 0.7. The CDRs were computed to glaucoma assessment where was reach sensitivity above 0.8, specificity of 0.7, F1-Score around 0.7 and AUC above 0.85. Finally, conclusions of segmentation methods showing adequate performance to be used in practical glaucoma screening.

Abstract
In a multimodal world the contact time between the teacher and the students is not always sufficient to ensure the effectiveness of the learning process. For the assimilation of concepts, students often endeavor on a search for the materials that best suit their learning needs. With the application of new technologies in teaching, study materials and support platforms are increasingly abundant and diverse. Additionally, recommendation algorithms overwhelm students with several options, sometimes hard to resist and select, especially after the COVID-19 restrictions, where the amount of connected time as increased. In this context, it is important for the teacher, to know which methods and materials the students use when they are autonomously developing their knowledge and skills. A survey was conducted within a group of engineering students at a Portuguese higher education institution with the main goal of characterizing the study habits and the materials that students. The obtained results are here reported and analyzed and compared with previous results from pre-pandemic study.

Abstract
Reliable estimations of parameter values and associated uncertainties are crucial for crop model applications in agro-environmental research. However, estimating many parameters simultaneously for different types of response variables is difficult. This becomes more complicated for grapevines with different phenotypes between varieties and training systems. Our study aims to evaluate how a standard least square approach can be used to calibrate a complex grapevine model for simulating both the phenology (flowering and harvest date) and yield of four different variety-training systems in the Douro Demarcated Region, northern Portugal. An objective function is defined to search for the best-fit parameters that result in the minimum value of the unweighted sum of the normalized Root Mean Squared Error (nRMSE) of the studied variables. Parameter uncertainties are estimated as how a given parameter value can determine the total prediction variability caused by variations in the other parameter combinations. The results indicate that the best-estimated parameters show a satisfactory predictive performance, with a mean bias of -2 to 4 days for phenology and -232 to 159 kg/ha for yield. The corresponding variance in the observed data was generally well reproduced, except for one occasion. These parameters are a good trade-off to achieve results close to the best possible fit of each response variable. No parameter combinations can achieve minimum errors simultaneously for phenology and yield, where the best fit to one variable can lead to a poor fit to another. The proposed parameter uncertainty analysis is particularly useful to select the best-fit parameter values when several choices with equal performance occur. A global sensitivity analysis is applied where the fruit-setting parameters are identified as key determinants for yield simulations. Overall, the approach (including uncertainty analysis) is relatively simple and straightforward without specific pre-conditions (e.g., model continuity), which can be easily applied for other models and crops. However, a challenge has been identified, which is associated with the appropriate assumption of the model errors, where a combination of various calibration approaches might be essential to have a more robust parameter estimation.

Abstract
When the COVID-19 pandemic hits Portugal in early March 2020, medical doctors, engineers and researchers, with the encouragement of the Northern Region Health Administration, teamed up to develop and build, locally and in a short time, a ventilator that might eventually be used in extreme emergency situations in the hospitals of northern Portugal. This letter tells you the story of Pneuma, a low-cost emergency ventilator designed and built under harsh isolation constraints, that gave birth to derivative designs in Brazil and Morocco, has been industrialized with 200 units being produced, and is now looking forward to the certification as a medical device that will possibly support a go-tomarket launch. Open intellectual property (IP), multi disciplinarity teamwork, fast prototyping and product engineering have shortened to a few months an otherwise quite longer idea-to-product route, clearly demonstrating that when scientific and engineering knowledge hold hands great challenges can be successfully faced.

Abstract
Monocular Visual Odometry techniques represent a challenging and appealing research area in robotics navigation field. The use of a single camera to track robot motion is a hardware-cheap solution. In this context, there are few Visual Odometry methods on the literature that estimate robot pose accurately using a single camera without any other source of information. The use of omnidirectional cameras in this field is still not consensual. Many works show that for outdoor environments the use of them does represent an improvement compared with the use of conventional perspective cameras. Besides that, in this work we propose an open-source monocular omnidirectional version of the state-of-the-art method Libviso2 that outperforms the original one even in outdoor scenes. This approach is suitable for central dioptric omnidirectional cameras and takes advantage of their wider field of view to calculate the robot motion with a really positive performance on the context of monocular Visual Odometry. We also propose a novel approach to calculate the scale factor that uses matches between laser measures and 3-D triangulated feature points to do so. The novelty of this work consists in the association of the laser ranges with the features on the omnidirectional image. Results were generate using three open-source datasets built in-house showing that our unified system largely outperforms the original monocular version of Libviso2.