Research & Projects
My work spans biomedical machine learning, biometrics, computer vision, physiological-signal processing, and reproducible research software.
ECG Biometrics Benchmarking
ECG-biometrics-bench is a research framework for reproducible benchmarking of ECG-biometric systems. The work focuses on standardized experimental workflows, heterogeneous physiological-signal datasets, biometric evaluation protocols, preprocessing, model training, and systematic comparison.
The associated solo-authored paper, “ECG-biometrics-bench: A Unified Framework for Reproducible Benchmarking of ECG Biometrics,” is available as a 2026 arXiv preprint and is currently under review.
Iris Recognition Framework
During my research at the University of Milan, I developed a modular Python framework for iris recognition that integrated more than 30 classical and deep-learning methods behind reusable interfaces. The framework was designed for systematic experimentation, comparison, and reproducibility across multiple stages of iris recognition, including work with ocular images captured using conventional cameras under uncontrolled conditions.
Ocular Computer Vision and Pupillometry
My current research collaboration at Vrije Universiteit Brussel involves computer vision for ocular analysis from RGB video. The work considers eye-region analysis, pupil localization and size estimation, temporal tracking, gaze-related measurements, and reliability under changes in illumination, viewing conditions, head movement, and image quality.
Motor-Imagery EEG and Brain-Computer Interfaces
My M.Sc. research focused on classification of motor-imagery EEG signals using deep-learning and signal-processing methods. This work included physiological-signal preprocessing, feature extraction, model development, and comparative evaluation.
A related conference paper investigated transfer-learning-based motor-imagery classification using convolutional neural networks.
Real-Time Road-Scene Computer Vision
In applied computer-vision work with the Road Traffic Injury Research Center, I worked on roadside-video pipelines for vehicle detection, temporal tracking, speed estimation, and inter-vehicle distance measurement.
The project involved data collection and annotation, preprocessing, model development, validation, robustness analysis, and operation under challenging conditions including illumination changes, rain, shadows, and partial occlusion.
Research Software
Across these projects, I place particular emphasis on:
- reproducible experimental design;
- modular Python research frameworks;
- systematic model evaluation;
- dataset and preprocessing consistency;
- testing and technical documentation;
- workflows that can be reproduced, compared, and extended by other researchers.