Luca Maria Di Giacomo
Combining Vision-Language Models and Language Models for Semantic Voice-Driven Teleportation in Virtual Reality.
Rel. Fabrizio Lamberti, Davide Calandra, Filippo Gabriele Prattico'. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering), 2026
|
|
PDF (Tesi_di_laurea)
- Tesi
Accesso limitato a: Solo utenti staff fino al 2 Ottobre 2029 (data di embargo). Licenza: Creative Commons Attribution Non-commercial No Derivatives. Download (16MB) |
Abstract
Virtual Reality (VR) environments require interaction techniques that are increasingly accessible, intuitive, and minimally dependent on handheld controllers. In locomotion tasks, traditional methods often rely on physical input devices, predefined targets, or direct geometric selection, limiting the possibility of expressing navigation goals in semantic terms. This thesis investigates semantic teleportation in Virtual Environments (VEs), enabling users to navigate through natural language commands that refer to objects, places, functional properties, or contextual relations within the scene. The proposed framework is based on semantic sensing points and precomputed scene representations. At development time, the virtual environment is analyzed from multiple viewpoints to collect visual observations of the scene.
These observations are processed by a Vision-Language Model (VLM) to extract semantic and spatial information, which is then integrated into a structured scene representation
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Corso di laurea
Classe di laurea
URI
![]() |
Modifica (riservato agli operatori) |
