Abstract
The environmental impact on the energy sector has become a significant concern, necessitating the implementation of Home Energy Management Systems (HEMS) to enhance the energy efficiency of buildings, reduce costs and greenhouse gas emissions, and ensure user comfort. This paper presents a novel approach to provide optimal day-ahead energy management plans in smart homes with Photovoltaic/Thermal (PVT) systems, aiming to achieve a balance between energy cost and user comfort. This multi-objective problem employs the Non-dominated Sorting Genetic Algorithm III as the optimization algorithm and the Nonlinear Auto-regressive with External Input to forecast the day-ahead meteorological variables, which serve as inputs to predict the PVT electrical and heat production in the thermal resistance model. The HEMS benefits from the time-of-use tariff due to the flexibility provided by the energy storage from a battery bank and a boiler. Furthermore, it performs a load scheduling for 10 controllable loads based on three feature parameters to characterize occupant behavior. A study case analysis revealed a cost reduction of approximately 66% in the solution close to the knee of the Pareto curve (S3 solution). The environmental impact on the energy sector has become a The PVT heat production was sufficient to meet the thermal demand of the showers. The proposed hybrid battery management model effectively eliminated the export of electricity to the grid, reducing consumption during peak periods and the maximum peak demand.

Abstract
With the increasing popularity of social media platforms like Instagram, there is a growing need for effective methods to detect and analyze abnormal actions in user-generated content. Deep learning is part of a broader family of machine learning methods based on artificial neural networks with representation learning that can learn complex patterns. This article proposes a novel deep learning approach for detecting abnormal actions in social media clips, focusing on behavioural change analysis. The approach uses a combination of Deep Learning and textural, statistical, and edge features for semantic action detection in video clips. The local gradient of video frames, time difference, and Sobel and Canny edge detectors are among the operators used in the proposed method. The method was evaluated on a large dataset of Instagram and Telegram clips and demonstrated its effectiveness in detecting abnormal actions with about 86% of accuracy. The results demonstrate the applicability of deep learning-based systems in detecting abnormal actions in social media clips.

Abstract
Small living spaces require ingenious solutions that are functional, ergonomic and, above all, reconfigurable. This project for smart, ergonomic and adjustable furniture was embraced by a team of students from different countries, universities and study areas enrolled in the European Project Semester (EPS) at Instituto Superior de Engenharia do Porto (ISEP). EPS is a design project where international students work in teams to create a solution to a real problem from scratch, analysing the state of the art, the market and the associated ethical and sustainability issues. As a project-based learning process, EPS aims to prepare engineering students to work together in multidisciplinary teams, develop personal skills and address the challenges of the contemporary world. The current project aims to design, simulate and test an ethically and sustainability-driven safe and transformable furniture. Amplea is the adjustable furniture solution developed by five EPS students in spring 2023. It transforms into a kitchen counter, dining table or standing desk. By transforming easily, Amplea’s design provides more comfort and saves space in small living spaces. This paper summarises the research, the design of the solution and the development and testing of the proof-of-concept prototype. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

Abstract
The high rates of mortality provoked by cardiovascular disorders (CVDs) have been rated by the OMS in the top among non-communicable diseases, killing about 18 million people annually. It is crucial to detect arrhythmias or cardiovascular events in an early way. For that purpose, novel portable acquisition devices have allowed long-term electrocardiographic (ECG) recording, being the most common way to discover arrhythmias of a random nature such as atrial fibrillation (AF). Nonetheless, the acquisition environment can distort or even destroy the ECG recordings, hindering the proper diagnosis of CVDs. Thus, it is necessary to assess the ECG signal quality in an automatic way. The proposed approach exploits the feature and meta-feature extraction of 5-s ECG segments with the ability of machine learning classifiers to discern between high- and low-quality ECG segments. Three different approaches were tested, reaching values of accuracy close to 83% using the original feature set and improving up to 90% when all the available meta-features were utilized. Moreover, within the high-quality group, the segments belonging to the AF class outperformed around 7% until a rate over 85% when the meta-features set was used. The extraction of meta-features improves the accuracy even when a subset of meta-features is selected from the whole set.

Abstract
Since digital media has become increasingly popular, video processing has expanded in recent years. Video processing systems require high levels of processing, which is one of the challenges in this field. Various approaches, such as hardware upgrades, algorithmic optimizations, and removing unnecessary information, have been suggested to solve this problem. This study proposes a video saliency map based method that identifies the critical parts of the video and improves the system's overall performance. Using an image registration algorithm, the proposed method first removes the camera's motion. Subsequently, each video frame's color, edge, and gradient information are used to obtain a spatial saliency map. Combining spatial saliency with motion information derived from optical flow and color-based segmentation can produce a saliency map containing both motion and spatial data. A nonlinear function is suggested to properly combine the temporal and spatial saliency maps, which was optimized using a multi-objective genetic algorithm. The proposed saliency map method was added as a preprocessing step in several Human Action Recognition (HAR) systems based on deep learning, and its performance was evaluated. Furthermore, the proposed method was compared with similar methods based on saliency maps, and the superiority of the proposed method was confirmed. The results show that the proposed method can improve HAR efficiency by up to 6.5% relative to HAR methods with no preprocessing step and 3.9% compared to the HAR method containing a temporal saliency map.

Abstract
In recent years, social media platforms have become an essential source of information. Therefore, with their increasing popularity, there is a growing need for effective methods for detecting and analyzing their content in real time. Deep learning is a machine learning technique that teaches computers to understand complex patterns. Deep learning techniques are promising for analyzing acoustic signals from social media posts. In this article, a novel deep learning approach is proposed for socially contextualized event detection based on acoustic signals. The approach integrates the power of deep learning and meaningful features such as Mel frequency cepstral coefficients. To evaluate the effectiveness of the proposed method, it was applied to a real dataset collected from social protests in Iran. The results show that the proposed system can find a protester's clip with an accuracy of approximately 82.57%. Thus, the proposed approach has the potential to significantly improve the accuracy of systems for filtering social media posts.