Community
Uno dei segni più tangibili dell'innovazione è il numero di pubblicazioni, parte del bagaglio delle esperienze accademiche ed industriali di dipendenti e collaboratori di synbrAIn. Insieme alle soluzioni, questo è uno degli asset più significativi di synbrAIn, che hanno permesso l'instaurarsi di numerose partnership, aziendali e non.
A CNN designed to detect myocardial scar tissue in cardiac slices by analyzing different combinations of parametric images derived from cine cardiac MRI.
To evaluate the diagnostic performance of an AI algorithm for the detection of acute appendicular fractures in the pediatric population on conventional CXR.
An investigation of complexity in time series of complex systems using entropy rate and conditional entropy.
Testing the diagnostic performance of an AI system in detecting COVID-19 pneumonia and typical bacterial pneumonia in patients who underwent a chest X-ray.
A survey on the deep learning technologies which, combined with wearable sensors, could help in diagnosing and monitoring PD.
A technique to exploit domain shift during training for medical image 3D segmentation, so that the additional data becomes more usable.
This study describes a computerized support system for microcirculation analysis based on a deep learning approach.
Through the means of Reinforcement Learning, this study explores the possibility of modern AIs to learn during the game.
Investigating the impact of various factors such as language, gender, and recording conditions on the effectiveness of different vocal tasks.
Discover a novel algorithm that uses Mask-RCNN to segment teeth in panoramic dental X-rays with high accuracy and Dice score.
This paper deals with the automatic detection of Myotonia from a task based on the sudden opening of the hand.
A model based on the temporal evolution of speech attractors in the reconstructed phase space to identify hallmarks of PD identification and progression.
This review aims at identifying the most widely employed and promising machine learning methodologies to analyse Parkinson's disease patients voice.
In this work, optical, differential air-pressure and acceleration signals, acquired by a chest-worn sensor, are elaborated to feed a deep network.
Following the current state-of-the-art, several ML pipelines were compared, and deep learning was also explored with a custom CNN architecture.
The aim of the study is to show whether it is possible to predict infectious disease outbreaks early, by using machine learning.
This work describes a Clinical Decision Support System for supporting Coeliac Disease diagnosis, based on a neural-network-based fuzzy classifier.
Analysis of the resistance to high levodopa doses of FoG, posture, speech, and altered gait features presenting in daily-ON therapeutic condition.
This work presents the design of a classifier to recognize human emotions from body gestures.
In this work, the influence of obesity and Gastro- Esophageal Reflux Disease (GERD) on voice is assessed, using a machine leargnin approach.
This study aims at assessing the diagnostic accuracy of minimally invasive point-of-care tests, which may eliminate the need for intestinal biopsy.
This study proposes a method for automatically detecting RBD from single-channel EEG data, by analysing segments recorded during both REM sleep and SWS.
This study assesses the feasibility of a telemedicine system for the evaluation of sleep quality through brief vocal recordings.
This study analyses the transition regions of specific phonetic groups to model the loss of motor control and the difficulties in movements in PD patients.
This work presents an algorithm to detect Parkinson's disease based on patients' speech data.
This study investigates the role of landmarks labeling before 3D facial acquisition, by comparing several measurements.
An intelligent system that visitors can use to automatically get a description of the scenes shown in a painted wooden ceiling.
The aim of this study is to provide a fluid-dynamic and biological description of unstable and stable (SA) plaques, according to OCT analysis.
This work explains how to exploit a cognitive architecture to model the characters of a story in an interactive storytelling system, accessible a humanoid robot
This work explores how children interact with touchless-enabled interactive displays, comparing their behavior with their chronological age.
This database represents the HNSCC largest available repository and will enable to develop decision support systems and machine learning models.
A novel methodology for 3D reconstruction of coronary artery bifurcations based on the integration of angiography with optical coherence tomography (OCT).
This study aims at developing a MRI-based radiomic signature as a prognostic marker for different clinical endpoints in NPC patients from non-endemic areas.
This work describes a robotic storytelling system, where the characters have been modelled as cognitive agents embodied in Pepper and NAO robots.
This work describes a methodology for developing an MRI-based radiomic signature for prognosis of overall survival in nasopharyngeal cancer.
This work shows how to build predictor models to forecast users’ interaction duration and distance when interacting via touchless mid-air gestures in public.
The aim of this study is to provide a reliable tool for the automated assessment of postural instability.
This study investigated the added value of end-point specific radiomic signatures (RS) in patients with primary ESTS and RPS.
Data suggest that radiomics can be used to predict outcomes in SCCHN. Here a radiomic model was developed to identify high and low risk of disease recurrence.
Two multi-metric approaches are proposed for image registration of brain images, combining mutual information and normalized gradient field filter.
The aim of this study is the application of a framework for the in-silico analysis of the disrupted hemodynamics due to an ulcerated lesion.
This work compare the performance of several radiomics‐based predictive models of response to induction chemotherapy (IC) in sinonasal cancers (SNCs).
This work describes an interactive storytelling system, accessible through the SoftBank robotic platforms NAO and Pepper.
This paper describes a 3D reconstruction method based on segmentation of patients' CT images with calcification, metallic artifacts and thrombus removal.
This work uses a virtual phantom to identify a set of radiomic features MRI brain images, which is stable to variations in image acquisition parameters.
In this work, the effect of time of repetition (TR) and time of echo (TE) on radiomic features was evaluated using a virtual phantom.
In this work, we report on a usability study carried out on a visual tool for patent infringement detection with 21 professional designers.
This study investigates the difference in terms of usability, effectiveness, and enjoyment perceived by users with ASD between touch and touchless interaction.
This work studies children's interactions with large display via touchless avatar-based interface, investigating the impact of interactiong on learning.
This work describes a touchless gesture elicitation and usability study to understand how to perform zoom actions while interacting with desktop displays.
This framework provides a web API to access and process data from coeliac patients, to be used also as a basis for diagnostic decision support systems.
An overview of novel areas of interest in pervasive displays research, based on the works presented during the 7th ACM Intl. Symposium on Pervasive Displays.
In this work, we present and compare two visual approaches to mid-air finger-based menu control in virtual reality environments.
This study validates a segmentation algorithm for the detection of both lumen contours and polymeric bioresorbable scaffold struts from 8-bit OCT images.
This work consists in the development of a prognostic model in oral cavity squamous cell carcinoma based using genomics features and MRI radiomics features.
This study identifies a set of radiomic features to predict the outcome of induction chemotherapy in sinonasal cancers.
This work presents a reconstruction methodology to compare patient-specific hemodynamics and neo-intimal thickening at nine months from the intervention.
This work introduces the idea of a sound-based system to overcome the display blindness, and some experiments to test its effectiveness.
This study investigates if avatar-based touchless interfaces can foster interest towards artworks, making them more accessible for people affected by ASD.
This work presents a novel technology to create controllable interactive displays, by exploiting thigmonastic behavior of plants such as Mimosa pudica.
This work describes a new way to assess stability and discrimination capacity of radiomic features.
I sistemi touchless possono favorire lʹindipendenza di pazienti a mobilità ridotta nellʹaccesso ai sistemi informativi, e quindi la prevenzione terziaria.
This work describes an information provision system allowing for touchless gestural interactions, along with a trial implementation within a University campus.
This word exploits the correlation between user's affective state and the simultaneous body expressions, to automatically recognize emotions from gestures.
In this study, an OCT-based reconstruction method was developed for the execution of CFD simulations of patient-specific coronary artery models.
KIND‐DAMA is a modular middleware for easing the development of interactive applications based on gestural input, applicable in a plethora of scenarios.
Aim of this study was to assess radiomic features stability and relevance for the analysis of medical images of soft-tissue sarcoma.
This work presents a system able to recognize human body gestures implementing a constrained training set reduction technique, allowing for real-time execution.
This work describes a system for conveying audience emotions during live musical exhibitions, controlling a humanoid robot based on mobile apps.
This work presents a brand new method for centerline extraction of vascular trees, by using computational fluid dynamics (CFD).
This work compares two versions of the same touchless gestural interface, to investigate the effect of user's Avatar on user experience and usability.
This work investigates the influence of a passive audience on the engagement of people with an interactive touchless gestural public display.
This work presents a persuasive pervasive system aimed at influencing users' behavior for reducing energy consumption in buildings.
In this study, a fully automatic method was developed for detection of both vessel contours and stent struts in tomography images.
The purpose of this paper is to present a modular middleware for gestural data and devices management.
We believe that the most promising killer application of Ubiquitous Computing is already here: the Human-to-Human Interaction mediated by computers.
In this paper describes a method to detect emotions from gestures using gestural data and a textual description of their meaning.
This paper describes a solution for an innovative and multimodal exploration of ancient book contents, using both touch and touchless gestures.
This paper describes a comparison study aimed at understanding how avatar-based touchless interface affect user experience and usability of interactive displays
Una panoramica sulle principali tecniche utilizzate per riconoscere i gesti, sfruttando i dati ottenibili dai dispositivi “Kinect-like”.
This work presents two challenges related to touchless networked displays: overcoming interaction blindness and performing evaluations in-the-wild
This work provides an overview of the main challenges in evaluating and designing touchless gestures in-the-wild, using novel user-centred design methods.
This work describes an approach that combines several image processing techniques with the goal of extracting principal palmprint lines.
This work describes a novel method to reproduce in real time the opening and closing gestures of a human hand, using data gathered from Kinect-like devices.
This paper describes how to use neural networks to detect in real time hand poses, based on data gathered from a Microsoft Kinect RGB-D sensor.
Le collaborazioni di synbrAIn sono numerose ed in costante evoluzione. I nostri partner ci hanno permesso di mettere in pratica le potenzialità dell'intelligenza artificiale in svariati ambiti applicativi, dal marketing alla ricerca, passando per l'healthcare e la gestione delle relazioni con i clienti (CRM).
