Collaborations

Università degli Studi di Siena, Italy

Within the collaboration, with the Visual Information Processing and Protection Group, directed by Prof. Mauro Barni, the research activities focused on the study and realization of innovative pattern recognition algorithms based on signal processing and artificial intelligence, in particular for biometric applications. In this context, the research included the study of original image processing methodologies for iris images based on biometric techniques for feature extraction and identity comparison, and on artificial intelligence models based on Deep Learning and Generative Adversarial Networks. The realized method enabled to perform the biometric anonymization of the iris in high-resolution images collected from the internet, while guaranteeing a high visual realism of the anonymized images.

STMicroelectronics, Catania, Italy

Within the collaboration, with Dr. Francesco Rundo, the research activities focused on the study and realization of innovative pattern recognition approaches based on signal and image processing and on artificial intelligence techniques, with a specific focus on autonomous driving applications. In this context, the research focused on original Deep Learning methods based on Convolutional Neural Networks for the semantic segmentation of images captured by a vehicle-mounted camera, and subsequently for the estimation of the distance of objects in the scene without requiring additional sensors.

Yildiz Technical University, Istanbul, Turkey

Within the collaboration, with Prof. Tulay Yildirim, the research activities focused on the analysis of recent cybersecurity methods, with a specific attention to artificial intelligence techniques using Deep Learning for intrusion detection in network communication scenarios. Furthermore, the research included the study of the databases most widely used in the intrusion detection field, with the purpose of highlighting their utility in training intelligent models and their future usefulness in comparion the performance of Deep Learning-based methods.

Carnegie Mellon University, Pittsburgh, PA, USA

Within the collaboration, with the Intelligent Sensor, Measurement, and Control Laboratory, directed by Prof. Mel Siegel, and with the Visual Intelligence Studio, directed by Prof. Yang Cai, the research activities focused on the study of recent methods for Ambient Intelligence based on last-generation sensors, innovative and adaptive communication infrastructures, and privacy-aware data processing techniques.

University of Toronto, ON, Canada

He was Visiting Researcher at the University of Toronto, ON, Canada, from June 1, 2017 to August 31, 2017. In this period, in collaboration with the Multimedia Laboratory, directed by Prof. Konstantinos N. Plataniotis, the research activities focused on the study and realization of innovative pattern recognition methods and algorithms based on image processing and computational intelligence techniques, with a specific focus on biometric applications. In this context, the research focused on innovative methods based on Convolutional Neural Networks with unsupervised training for the adaptive analysis of palmprint samples captured using contactless and non ideal acquisition procedures. The major contributions consist in the use of Convolutional Neural Networks able to extract highly-discriminative biometric information, while being trained with unlabeled biometric data.
Subsequently, he was Visiting Researcher at the University of Toronto, ON, Canada, from December 2, 2019 to March 2, 2020. In this period, the research focused on the study and realization of innovative pattern recognition methods and algorithms based on image processing and computational intelligence, with a specific attention to biomedical applications.

Università degli Studi di Salerno, Fisciano, Italy

The research activities were performed in collaboration within the project Contactless Multibiometric Mobile System in the Wild (COSMOS), funded by the MIUR. With the Biometric and Image Processing Lab, directed by Prof. Michele Nappi, the research activities focused on the study and realization of innovative pattern recognition methods and algorithms based on signal processing and computational intelligence. In particular, the research focused on feature extraction methods applied on biometric samples captured in heterogeneous conditions and on information fusion techniques for the recognition of fingerprints, palmprint, and face. The methods have been validated with the application on biometric samples captured using contactless and non-idea acquisition procedures, enabling to perform the recognition in a less-constrained way, with greater usability and social acceptance than the methods currently used.

Universidad Rey Juan Carlos, Madrid, Spain

The research activities were performed in collaboration within the project Automated Border Control Gates for Europe (ABC4EU), funded by the European Commission. With the Face Recognition and Artificial Vision Research Laboratory for Advanced Security, directed by Prof. Enrique Cabello Pardos, the research activities focused on the study and realization of innovative pattern recognition methods and algorithms based on image processing and computational intelligence. In particular, the research focused on novel techniques for highly-usable biometric recognition, by studying and realizing innovative methods for the fusion of multi-sensor signals and images. The realized methods have been validated in the context of the high-security, multi-biometric recognition techniques present in recent automated border control systems.

European Centre for Soft Computing, Mieres, Spain

The research activities were performed in collaboration within the project I-PAN - Innovative Poplar Low Density Structural Panel, funded by the European Commission. With the Laboratory of Applications of Fuzzy Logic and Evolutionary Algorithms, directed by Prof. Sergio Damas Arroyo, the research activities focused on the study and realization of innovative pattern recognition methods and algorithms based on image processing and computational intelligence. In particular, the research focused on industrial application scenarios, by studying and realizing innovative methods for the granulometric analysis of particles captured using less-constrained acquisition procedures, in the presence of high levels of noise and with frequent occlusions. The realized methods have been validated in the context of the industrial production of panels made with wood strands.