Researcher in Computer Vision for Sports Analytics.
Ph.D. student specializing in Video Understanding,
Multimodal Learning, and AI-driven Analytics.
5VREAL – 5G Volley Reality Experience and Analytics Live
The Free University of Bolzano, EMG Italy, Fondazione Bruno Kessler, Small Pixels, and Vodafone Italy successfully collaborated to transform the analysis, experience, and broadcasting of volleyball through cutting-edge technologies. By leveraging 5G, Artificial Intelligence, and Video Analytics Systems, the project redefined how sports are observed and understood in real time.
Gate-Shift-Pose: A sport-tailored, pose-enhanced action recognition framework
Gate-Shift-Pose (GSP) is a multimodal architecture designed to enhance action recognition in sports by integrating skeleton-based pose information with RGB frames. The model builds on the Gate-Shift-Fuse (GSF) network and introduces early- and late-fusion strategies to better capture the intricate motion dynamics of sports.
Real-Time Object Detection with Meta Aria Glasses and YOLO
This project enables real-time object detection using different versions of YOLO (You Only Look Once) in conjunction with the Meta Project Aria Glasses. It allows streaming video input via USB or WiFi and performs object detection on the streamed data.
Transformers-Based Fire Detection in Forest Environments
Computer Vision project focused on detecting smoke and fire in wild environments. The Google Vision Transformer was fine-tuned on a custom dataset.
GPT-Based Vocal Virtual Assistant
GIVA is a vocal assistant that combines speech recognition and text-to-speech with the capabilities of GPT. Prompts are engineered so that GPT provides outputs that are short and adapted to be converted to audio.
T5-Based Headline Generator
Text-2-Title is a T5-based model fine-tuned for headline generation. The model was trained on over 110K article abstracts and related titles.
Toxic Comments Classification
NLP project focused on the Semantic Representation of text. Different techniques are proposed and compared, and a final web app classifies comments on demand.
Theme Park Ride Accidents Analysis
Data Integration project focused on theme park accident analysis leveraging Semantic Web and Linked Data technologies. A web app with interactive visualizations allows the exploration of the results.
Real-Time Credit Card Transactions Analysis
Big Data analysis project focused on transactional data stream processing. A real- time dashboard presents results and insights.
LangChain for LLM Application Development
Building Systems with the ChatGPT API
Azure AI Fundamentals
