In recent years, the emergence of organoid-related technology has transformed the landscape of biomedical research by providing near-physiological models that closely mimic human tissue. Intestinal organoids are three-dimensional structures derived from intestinal stem cells state that offer new potential for disease modeling, drug testing, and personalized medicine. However, the complexity of these heterogeneous models requires innovative solutions to optimize the organoids’ characterization and monitoring. From a technological perspective, the automated analysis of intestinal organoid images is essential for high-throughput screening, yet remains challenging due to morphological variability and imaging conditions. To overcome these challenges, this work proposes DECSEFE-Org, a hierarchical and modular framework that combines AI-based models for real-time organoid detection (YOLOv7), classification (DenseNet169), and segmentation (SAM). The pipeline further includes feature extraction to quantify morphological parameters and explainable AI (XAI) modules based on SHAP to support biological interpretation in decision-making processes. The framework was tested on a publicly labeled organoids dataset and achieved 86.5% accuracy in classification and 0.92 Dice score in segmentation. By using DECSEFE-Org is possible to accelerate drug response studies and improve treatment efficacy evaluation, overcoming the speed and accuracy of traditional methods. Experimental validation demonstrates the pipeline’s ability to provide near-real-time results, scalable analysis with low-latency in detection and segmentation, while maintaining interpretability and high-throughput suitability, opening the way for automated and scalable analysis of organoids in biomedical research.

Urban traffic is one of the most important issues for smart cities, and real-time management is critical for improving mobility and reducing congestion. Traditionally, classification methods based on machine learning need frequent retraining, which is inefficient for adaptive traffic management systems. This work introduces a novel hybrid approach that combines Incremental Learning (IL) with Federated Learning (FL) techniques to support continuous model adaptation without centralizing data or restarting the training process from scratch. The proposed approach employs Convolutional Neural Networks (CNNs) to classify traffic conditions from junction camera feeds and includes a specific mechanism to mitigate the Catastrophic Forgetting (CF) issue, a common drawback in IL. This solution enhances model performance in a decentralized way while protecting data privacy and encouraging knowledge sharing across distributed nodes. Experimental results on a publicly available dataset reveal that this approach significantly improves dynamic traffic management, achieving over 96% validation accuracy throughout the IL process for multiple clients, with minimal loss and no need to centralize data. This work sets the basis for a more effective and secure traffic management infrastructure in smart cities.

Organoids replicate key aspects of organ function and present a microcosmic representation of the in-vivo state, offering unprecedented opportunities for disease modeling, drug testing, and personalized medicine. However, the complexity of these heterogeneous biological models requires innovative approaches to optimize the analysis and monitoring. This work outlines an Intelligent Cyber-Biological System (ICBS) with a multi-layered MAPE-K structure for enhanced analysis and real-time organoid monitoring. Starting with the Laboratory Layer, the ICBS harnesses real-time data from organoid cultures, which the Data Layer processes and archives for insight extraction. The Intelligence Layer employs advanced machine learning (ML) models for pattern recognition and predictive analytics, identifying critical development stages and treatment responses. In the Decision Layer, perceptions derived from ML-generated guide strategic decision-making for culture condition adjustments and further examinations. The ICBS framework engages researchers in its findings, fostering decision implementation and a responsive feedback system. Its real-world application in the intestinal organoid case study demonstrates the feasibility and effectiveness of the proposed solution, increasing the efficiency of experimental processes and the predictive accuracy of treatment outcomes.

Cardiovascular diseases are currently the major causes of death globally. Among the strategies to prevent cardiovascular issues, the automated classification of heart sound abnormalities is an efficient way to detect early signs of cardiac conditions leading to heart failure or other, even asymptomatic, complications, quite effective for timely interventions. Despite the significant improvements in this field, there are still limitations due to the lack of solutions, available data-sets and poor (mainly binary - normal vs abnormal) classification models and algorithms. This paper presents a Medical Cyber-Physical System (MCPS) for the automatic classification of heart valve diseases onsite, in a timely manner. The proposed MCPS, indeed, can be deployed into personal and mobile devices, addressing the limitations of existing solutions for patients, healthcare practitioners, and researchers, through an efficient and easy accessible tool. It combines different neural network models trained on a new Italian dataset of 132 adult patients covering 9 heart sound categories (1 normal and 8 abnormal), also validated against two main open-access (Physionet/CinC Challenge 2016 and Korean) datasets. The overall MCPS performance (time, processing and energy resource utilization) and the high accuracy of the models (up to 98%) demonstrated the feasibility of the proposed solution, even with few data. The dataset supporting the findings of this paper is available upon request to the authors.

Wastewater treatment is a critical process in urban and industrial settlements aiming to clean and protect the water as well as the overall environment. Wastewater management systems are conceived explicitly for purifying wastewater, providing clean water efficiently, but this is a hard task due to frequent and quite unpredictable fluctuations of inlet wastewater flows, arising from (random) rain water or (periodical, e.g. day-night) sewage sources, sometimes also leading to failures and outages. To ensure the quality of the clean water out above a threshold and keep the overall system operating, this paper proposes the smart wastewater intelligent management system (SWIMS). It monitors and controls inlet and outlet flows as well as the water quality and parts of the plant as a cyber-physical system (CPS), starting from an Environmental Internet of Things (EIoT) platform. The data generated from the treatment plant is collected in an information system hosted by a server together with an intelligent system that processes this information in a real-time fashion and provides the feedback for optimizing the plant to maintain a good quality of water over time. Such an intelligent system exploits deep learning approaches to control the behaviour of the wastewater treatment system through anomaly detection, supporting decision making on it. SWIMS has been implemented in a real case study deployed in Briatico, Italy. The data and results collected from such a case study are presented, analyzed and discussed in this paper, demonstrating the feasibility and the effectiveness of the SWIMS solution.

The primary goal of a wastewater treatment system is to take care of the environment as well as of people health by purifying sewage water. In urban and industrial environments, wastewater management is non-trivial since it has to deal with abnormal fluctuations in incoming water flows (due to rainwater or human and industrial sewage) that may cause failures and outages to the entire purification process. This paper proposes a solution based on a smart system to ensure the clean water quality by keeping the wastewater treatment system efficient. It is able to constantly and real time monitoring both the purity of the water and the inlet and outlet flows enforcing on them proper policies based on the monitored values thus acting as a cyber-physical system (CPS). The raw data, generated by an Environmental Internet of Things (EIoT) platform part of a real case study implemented in Briatico (Italy), is collected and hosted in a server that can process and manage real-time information about the plant.

The Internet of Medical Things (IoMT), combined with interconnected wearable devices and medical-grade sensors, can play an essential role in healthcare evolution. By exploiting the data generated by the plethora of interconnected devices (vital parameters, location-based info, patients activity and more), advanced ICT systems can be put in place with predicting capabilities. This way potentially critical situations, that may evolve in serious complications to patients’ well-being, can be promptly recognized and successfully addressed, first of all, to save lives and secondarily to limit economical damages. To support continuous patient monitoring in public and private healthcare, this paper proposes "DILoCC", an architecture to manage wearable devices, sensors and applications, that uses a Distributed Incremental Learning (DIL) approach to exploit cooperation among the sensing devices and increase the overall system efficiency through the mitigation of "Catastrophic Forgetting" consequences.

In this work, we propose a low-cost Smart and Healthy Intelligent Room System (SHIRS), able to monitor Indoor Air Quality (IAQ) by enhancing edge-based computation. SHIRS exploits the ability to run Machine Learning (ML) algorithms to infer humans presence (headcount) from environmental data analysis. Experimental results show the validity of the proposed approach, demonstrate the potential of edge-based computing and push towards the adoption of smart integrated Cloud-IoT frameworks for environmental monitoring and control.