Artificial intelligence (AI), and specifically deep learning (DL) models, are rapidly gaining traction in healthcare to analyze complex medical images and support clinical decision-making. However, DL models are often considered black boxes due to the lack of a clear explanation when providing predictions. Explainable artificial intelligence (XAI) methods are emerging as an effective way to make models explainable for developers and provide interpretable outputs for clinicians. This review presents a taxonomy of the most widely used XAI methods for image classification, with related benefits and drawbacks. Furthermore, it examines whether the type of classifier affects the choice of an explainability technique and investigates the impact of black boxes on the healthcare environment. The analysis considered papers published between January 2020 and July 2025 in Scopus and Google Scholar, utilizing the PRISMA guidelines to enhance reporting. Sixty-nine papers were identified as suitable for classifying XAI methods in four categories based on backpropagation, perturbation, attention, and concept. The results show increased use of backpropagation-based techniques, which offer simple and intuitive heatmaps. Perturbation-based methods are frequently employed to validate model robustness, but they are computationally expensive. Finally, concept-based and attention-based approaches are less widespread but represent a promising solution towards explanations that align with human semantics and reflect the intrinsic model behavior. Future research should focus on combined approaches and concept methods that generate explanations in the same semantic field as clinicians and are computationally suitable for healthcare environments, paving the way for transparent and clinically reliable DL systems.

Stress is a reaction that occurs when a person perceives, with or without awareness, an imbalance between requests and available resources. Relying on this definition, we have carried out an experiment in a Virtual Reality environment to elicit (light) stress in the user and analyze the emotional responses with electroencephalography (EEG). The virtual environment is divided in eight parts; in each of them a stressor has been put in action, meaning that in every part the participants perform a task, but a specific resource is missing (time, knowledge, control, salvation, no or too many alternatives, engagement, self-confidence). EEG is used to assess the emotional response with the aid of Valence/Arousal/Dominance/Stress indicators presented in previous literature. Nine indicators calculated for 87 participants, labeled according to self-assessment replies (post-experimental questionnaires), were classified with eXtreme Gradient Boosting , k-Nearest Neighbor, Support Vector Machine and Random Forest classifiers . The lowest results in terms of accuracy were obtained with k-Nearest Neighbor (around 70 %), whilst the highest ones were obtained with eXtreme Gradient Boosting and Random Forest (above 98 %), showing that EEG could be a valuable tool to assess the emotional response in stressful situations, with a particular focus on the Stress indicators.

Abstract The “Augmented Humans” term refers to the opportunity to improve human possibilities by using innovative technologies such as Artificial Intelligence (AI) and Extended Reality (XR). Digital therapies, particularly suitable for those treatments requiring multiple sessions, are increasingly being adopted for home-based treatment, enabling continuous monitoring and rehabilitation for patients, thus alleviating the burden on healthcare facilities by facilitating remote therapy sessions and follow-up visits. Among these, the Mirror Therapy (MT) for patients suffering from Phantom Limb Pain (PLP) could benefit greatly. This paper proposes a novel “Augmented Humans” framework for the treatment of PLP through home-based MT; the framework is designed to consider the activities carried on by the therapy center, the patient, and the system supporting the treatment. Moreover, an XR-based solution that integrates a Deep Learning (DL) approach has been developed to provide patients with a self-testing and self-assessment tool for conducting at-home rehabilitation sessions independently, even in the absence of physical medical staff. The DL algorithm enables real-time monitoring of rehabilitation exercises and automatic provision of personalized feedback on the gesture’s performance, supporting the progressive improvement of the patient’s movements and his ability to adhere to the treatment plan. The technical feasibility and usability of the proposed framework have been evaluated with 23 healthy subjects, highlighting an overall positive user experience. Remarkable results were obtained in terms of automatic gesture evaluation, with macro averaged accuracy and F1-score of 95%, paving the way for the adoption of the “Augmented Humans” approach in the healthcare domain.

Abstract In orthopedic surgery and maxillofacial there is a growing use of augmented reality (AR) as a technology to increase the visual perception of the surgeon in the operating room. The objective of this review is to analyze the state of the art in the use of AR for osteotomies, highlighting the advantages and the most-known open issues to be addressed in the future research. Scopus, Web of Science, Pubmed and IEEE Xplore databases have been explored with a keyword search, setting the time limits from January 2017 to January 2023, inclusive. Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines has been used in this review, focusing on anatomical districts, real-virtual environment interaction, advantaged and limitations of existing AR-based applications. 49 articles met the inclusion criteria and have been selected in the final analysis. For the sake of clarity, works have been grouped according to the anatomical district, but also the real-virtual environment interaction methodology was reported, as well as information regarding accuracy assessment. A Quality Function Deployment (QFD) has been used to assess the AR-based solutions with regards to the more traditional freehand (FH) and Patient Specific Template (PST) approaches. Finally, some suggestions to foster the AR-based solution adoption in osteotomies treatment have been drawn, considering the highlighted advantages and limitations of this technology. The AR resulted to meet the surgeons’ needs more than other traditional approaches. Among the emerged advantages, AR can lead to a better surgical field accessibility, more flexible solutions and lower the management effort. Nonetheless, future research should address some well-known issues, among which the calibration time, the robustness of the tracking, and the HMDs discomfort.

Abstract This study investigates the use of electroencephalography (EEG) to characterize emotions and provides insights into the consistency between self-reported and machine learning outcomes. Thirty participants engaged in five virtual reality environments designed to elicit specific emotions, while their brain activity was recorded. The participants self-assessed their ground truth emotional state in terms of Arousal and Valence through a Self-Assessment Manikin. Gradient Boosted Decision Tree was adopted as a classification algorithm to test the EEG feasibility in the characterization of emotional states. Distinctive patterns of neural activation corresponding to different levels of Valence and Arousal emerged, and a noteworthy correspondence between the outcomes of the self-assessments and the classifier suggested that EEG-based affective indicators can be successfully applied in emotional characterization, shedding light on the possibility of using them as ground truth measurements. These findings provide compelling evidence for the validity of EEG as a tool for emotion characterization and its contribution to a better understanding of emotional activation.

Abstract The 6D pose estimation of an object from an image is a central problem in many domains of Computer Vision (CV) and researchers have struggled with this issue for several years. Traditional pose estimation methods (1) leveraged on geometrical approaches, exploiting manually annotated local features, or (2) relied on 2D object representations from different points of view and their comparisons with the original image. The two methods mentioned above are also known as Feature-based and Template-based, respectively. With the diffusion of Deep Learning (DL), new Learning-based strategies have been introduced to achieve the 6D pose estimation, improving traditional methods by involving Convolutional Neural Networks (CNN). This review analyzed techniques belonging to different research fields and classified them into three main categories: Template-based methods, Feature-based methods, and Learning-Based methods. In recent years, the research mainly focused on Learning-based methods, which allow the training of a neural network tailored for a specific task. For this reason, most of the analyzed methods belong to this category, and they have been in turn classified into three sub-categories: Bounding box prediction and Perspective-n-Point (PnP) algorithm-based methods, Classification-based methods, and Regression-based methods. This review aims to provide a general overview of the latest 6D pose recovery methods to underline the pros and cons and highlight the best-performing techniques for each group. The main goal is to supply the readers with helpful guidelines for the implementation of performing applications even under challenging circumstances such as auto-occlusions, symmetries, occlusions between multiple objects, and bad lighting conditions.

Abstract Recently a wide variety of applications has been developed integrating 3D functionalities. Advantages given by the possibility of relying on depth information allows the developers to design new algorithms and to improve the existing ones. In particular, for what concerns face morphology, 3D has led to the possibility to obtain face depth maps highly close to reality and consequently an improvement of the starting point for further analysis such as Face Detection, Face Authentication, Face Identification and Face Expression Recognition. The development of the aforementioned applications would have been impossible without the progress of sensor technologies for obtaining 3D information. Several solutions have been adopted over time. In this paper, emphasis is put on passive stereoscopy, structured light, time-of-flight (ToF) and active stereoscopy, namely the most used technologies for the cameras design and fulfilment according to the literature. The aim of this article is to investigate facial applications and to examine 3D camera technologies to suggest some guidelines for addressing the correct choice of a 3D sensor according to the application that has to be developed.