Francesco Gentiluomo
An Integrated Vision-Based System for Gesture Interaction and Driver Monitoring in Smart Cockpits.
Rel. Davide Calandra. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering), 2026
|
Preview |
PDF (Tesi_di_laurea)
- Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives. Download (5MB) | Preview |
Abstract
The transition towards smart vehicles has driven a significant evolution in active safety systems, shifting the technological focus directly onto the driver. Within next-generation cabins (smart cockpits), attention monitoring technologies, known as Driver Monitoring Systems (DMS), coupled with the development of intelligent interfaces, have therefore become fundamental. In this context, this thesis proposes an integrated system that enables natural, touchless interaction via dynamic hand gestures while simultaneously monitoring the driver's attention state. The system integrates both functionalities (interaction and monitoring) using a single RGB-IR camera positioned at the level of the rearview mirror, thus optimizing hardware resources within the cabin. For gesture recognition, the features extracted via the Google MediaPipe framework feed a bidirectional LSTM neural network, trained on a custom dataset of 2160 samples (acquired in multiple configurations to increase variety).
Concurrently, the Driver Monitoring module leverages facial feature extraction to evaluate the driver's fatigue and gaze direction, operating in synergy with a YOLO model dedicated to detecting smartphones as a potential source of distraction
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Corso di laurea
Classe di laurea
Aziende collaboratrici
URI
![]() |
Modifica (riservato agli operatori) |
