Current medical practice for determining hemoglobin concentration (which is especially important for anemic patients in need of blood transfusion) involves frequent blood tests. In this work, we propose an alternative, non-invasive approach to hemoglobin estimation, based on image analysis of a specific conjunctival region. Our ultimate goal is to develop an easy-to-use wearable device that patients themselves can employ at home to autonomously assess their need of blood transfusion. In this paper, we detail the prototype of our device and the methodology for extracting key information from the color values of the acquired image. Tests conducted on 77 anemic and healthy patients show significant correlation between the real hemoglobin value obtained through blood sampling and the value estimated by our algorithm. A prototypical binary classification algorithm for assessing the need of blood transfusion yielded good results in terms of accuracy, specificity and sensitivity, thus making it possible to avoid a significant number of blood tests.

The COVID-19 pandemic marked an important breakthrough in human progress: from working habits to social life, the world population’s behaviours changed according to the new lifestyle requirements. In this changing environment, university courses and learning methods evolved along with other “remote” working activities. For this quasi-experimental study, we discuss the effectiveness of the changes made by the LUMSA University in Rome, comparing two different groups of students who attended a master’s course with blended and fully remote methodologies. Here, we focused our attention on the paradigm shift, comparing the data gathered during the blended course in the 2019/2020 academic year with data gathered during the same course, but conducted fully online, in the academic year 2020/2021. Considering the sample size and type, the group comparison was made using a non-parametric test (U-test). The statistical analysis results suggest that there was no substantial difference between the students’ performance, confirming that the course changes made to adapt to the pandemic situation were successful and that learning effectiveness was preserved.

Early warning systems (EWSs) are crucial for optimising predictive maintenance strategies, especially in the industrial sector, where machine failures often cause significant downtime and economic losses. This research details the creation and evaluation of an EWS that incorporates deep learning methods, particularly using Long Short-Term Memory (LSTM) networks enhanced with attention layers to predict critical machine faults. The proposed system is designed to process time-series data collected from an industrial printing machine’s embosser component, identifying error patterns that could lead to operational disruptions. The dataset was preprocessed through feature selection, normalisation, and time-series transformation. A multi-model classification strategy was adopted, with each LSTM-based model trained to detect a specific class of frequent errors. Experimental results show that the system can predict failure events up to 10 time units in advance, with the best-performing model achieving an AUROC of 0.93 and recall above 90%. Results indicate that the proposed approach successfully predicts failure events, demonstrating the potential of EWSs powered by deep learning for enhancing predictive maintenance strategies. By integrating artificial intelligence with real-time monitoring, this study highlights how intelligent EWSs can improve industrial efficiency, reduce unplanned downtime, and optimise maintenance operations.