Lorenzo Formentin
Engineering a Unified and Reproducible Python Framework for Multimodal Biosignal Feature Extraction in the MultiMotion Emotion Recognition Project.
Rel. Gabriella Olmo, Vito De Feo. 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 (7MB) |
Abstract
Automatic emotion recognition is a central challenge in affective computing, where artificial intelligence is used to infer emotional states objectively from physiological and behavioural signals. Such estimates are expected to reflect the affective experience reported by the individual when emotional awareness and reporting are sufficiently preserved. The MultiMotion project addresses this challenge through a multimodal framework that combines physiological and behavioural modalities, including Galvanic Skin Response (GSR), cardiac measures such as Heart Rate Variability (HRV), pupil size, and facial expression, with participant-specific subjective ratings used to derive continuous affective targets for Valence and Arousal prediction. Over the development of MultiMotion, feature-extraction methods for the different modalities were created independently, often at different stages of the project and using separate code bases and software environments.
As a result, the project accumulated a substantial collection of signal-processing and feature-extraction code without a common software architecture
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Corso di laurea
Classe di laurea
Aziende collaboratrici
URI
![]() |
Modifica (riservato agli operatori) |
