Abstract
This study focuses on investigating different CNN architectures and assessing their effectiveness in classifying Diabetic Retinopathy, a diabetes-associated disease that ranks among the primary causes of adult blindness. However, early detection can significantly prevent its debilitating consequences. While regular screening is advised for diabetic patients, limited access to specialized medical professionals can hinder its implementation. To address this challenge, deep learning techniques provide promising solutions, primarily through their application in the analysis of fundus retina images for diagnosis. Several CNN architectures, including MobileNetV2, VGG16, VGG19, InceptionV3, InceptionResNetV2, Xception, DenseNet121, ResNet50, ResNet50V2, and EfficientNet (ranging from EfficientNetB0 to EfficientNetB6), were implemented to assess and analyze their performance in classifying Diabetic Retinopathy. The dataset comprised 3662 Fundus retina images. Prior to training, the networks underwent pre-training using the ImageNet database, with a Gaussian filter applied to the images as a preprocessing step. As a result, the Efficient-Net stands out for achieving the best performance results with a good balance between model size and computational efficiency. By utilizing the EfficientNetB2 network, a model was trained with an accuracy of 85% and a screening capability of 98% for Diabetic Retinopathy. This model holds the potential to be implemented during the screening stages of Diabetic Retinopathy, aiding in the early identification of individuals at risk.

Abstract
The exponential growth in the digitisation of services implies the handling and storage of large volumes of data. Businesses and services see data sharing and crossing as an opportunity to improve and produce new business opportunities. The health sector is one area where this proves to be true, enabling better and more innovative treatments. Notwithstanding, this raises concerns regarding personal data being treated and processed. In this paper, we present a patient-centric platform for the secure sharing of health records by shifting the control over the data to the patient, therefore, providing a step further towards data sovereignty. Data sharing is performed only with the consent of the patient, allowing it to revoke access at any given time. Furthermore, we also provide a break-glass approach, resorting to Proxy Re-encryption (PRE) and the concept of a centralised trusted entity that possesses instant access to patients' medical records. Lastly, an analysis is made to assess the performance of the platform's key operations, and the impact that a PRE scheme has on those operations.

Abstract
Many clinical applications involve in-bed patient activity monitoring, from intensive care and neuro-critical infirmary, to semiology-based epileptic seizure diagnosis support or sleep monitoring at home, which require accurate recognition of in-bed movement actions from video streams. The major challenges of clinical application arise from the domain gap between common in-the-lab and clinical scenery (e.g. viewpoint, occlusions, out-of-domain actions), the requirement of minimally intrusive monitoring to already existing clinical practices (e.g. non-contact monitoring), and the significantly limited amount of labeled clinical action data available. Focusing on one of the most demanding in-bed clinical scenarios - semiology-based epileptic seizure classification - this review explores the challenges of video-based clinical in-bed monitoring, reviews video-based action recognition trends, monocular 3D MoCap, and semiology-based automated seizure classification approaches. Moreover, provides a guideline to take full advantage of transfer learning for in-bed action recognition for quantified, evidence-based clinical diagnosis support. The review suggests that an approach based on 3D MoCap and skeleton-based action recognition, strongly relying on transfer learning, could be advantageous for these clinical in-bed action recognition problems. However, these still face several challenges, such as spatio-temporal stability, occlusion handling, and robustness before realizing the full potential of this technology for routine clinical usage.

Abstract
Climate change, limited natural resources, and the increase in the world's population impose society to produce food more sustainably, with lower energy and water consumption. The use of robots in agriculture is one of the most promising solutions to change the paradigm of agricultural practices. Agricultural robots should be seen as a way to make jobs easier and lighter, and also a way for people who do not have agricultural skills to produce their food. The PixelCropRobot is a low-cost, open-source robot that can perform the processes of monitoring and watering plants in small gardens. This work proposes a mission supervisor for PixelCropRobot, and general agricultural robots, and presents a prototype of user interface to this mission supervision. The communication between the mission supervisor and the other components of the system is done using ROS2 and MQTT, and mission file standardized. The mission supervisor receives a prescription map, with information about the respective mission, and decomposes them into simple tasks. An A* algorithm then defines the priority of each mission that depends on factors like water requirements, and distance travelled. This concept of mission supervisor was deployed into the PixelCropRobot and was validated in real conditions, showing a enormous potential to be extended to other agricultural robots.

Abstract
This paper aims to study battery response under two operation strategies to analyze the annual cycles and operation costs (revenues) via sensitivity analysis. A battery model that considers performance parameters (AC-AC RTE, DOD, and C-rates) for different technologies is approached to identify how these parameters influence battery behavior and revenue. Strategies refer to (A) energy arbitrage, EA, and (B) EA and the provision of tertiary reserve. Simulations conducted for real data from Portuguese electricity and regulation markets showed regardless of the strategy used, the annual cycles and revenue are dominated by the performance parameters, instead of price volatility. In addition, for batteries with higher C-rates, as the AC-AC RTE is reduced up to 80%, the annual cycles and revenues are significantly reduced to 50% and 45% respectively, regarding its ideal model (100% AC-AC RTE). For lower C-rates, the annual cycles and revenues are slightly reduced with AC-AC RTE reductions. Specifically, strategy B revealed that annual cycles and revenue could also be influenced by the capacity requirements and the control area where batteries are providing services. © 2024 IEEE.

Abstract
Prostate cancer is one of the most common types of cancer in men. The ISUP grade and Gleason Score are terms related to the classification of this cancer based on the histological characteristics of the tissues examined in a biopsy. This paper explains an approach that utilizes and evaluates pre-trained models such as ResNet-50, VGG19, and InceptionV3, regarding their ability to automatically classify prostate cancer and its severity based on images and masks annotated with ISUP grades and Gleason Scores. At the end of the training, the performance of each trained model is presented, as well as the comparison between the original and predicted data. This comparison aims to understand if this approach can indeed be used for a more automated classification of prostate cancer.