Automatic detection of emotions has many applications, but there is the capacity for improvement with respect to level of automation, accuracy, and speed. A system for recognizing emotions through facial expressions in live video streams and video sequences is presented in this paper. We have implemented an Active shape Model (ASM) tracker, which tracks 116 facial landmarks via web-cam input. The tracked landmark points are used to extract face expression features. An Support Vector Machine (SVM) based classifier is implemented which gives rise to robust our system by recognizing seven expressions rather than only six expression as in the most of face expression systems. This technique is applied for the automated identification of the psychological state that exhibits a very strong correlation with the detected features.
Real time intention recognition / Anwar, Suzan; Milanova, Mariofanna; Bigazzi, Andrea; Bocchi, Leonardo; Guazzini, Andrea. - ELETTRONICO. - (2016), pp. 1021-1024. (Intervento presentato al convegno 42nd Conference of the Industrial Electronics Society, IECON 2016 tenutosi a Palazzo dei Congressi, ita nel 2016) [10.1109/IECON.2016.7794016].
Real time intention recognition
BOCCHI, LEONARDO;GUAZZINI, ANDREA
2016
Abstract
Automatic detection of emotions has many applications, but there is the capacity for improvement with respect to level of automation, accuracy, and speed. A system for recognizing emotions through facial expressions in live video streams and video sequences is presented in this paper. We have implemented an Active shape Model (ASM) tracker, which tracks 116 facial landmarks via web-cam input. The tracked landmark points are used to extract face expression features. An Support Vector Machine (SVM) based classifier is implemented which gives rise to robust our system by recognizing seven expressions rather than only six expression as in the most of face expression systems. This technique is applied for the automated identification of the psychological state that exhibits a very strong correlation with the detected features.File | Dimensione | Formato | |
---|---|---|---|
FinalPaperIECON2016.pdf
Accesso chiuso
Tipologia:
Pdf editoriale (Version of record)
Licenza:
Tutti i diritti riservati
Dimensione
4.3 MB
Formato
Adobe PDF
|
4.3 MB | Adobe PDF | Richiedi una copia |
I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.