A voir également
Profil
Titre : Doctorant.e
Ne fait plus partie de l'unité
----------------------------------------------------------------------------------------------------------------------------------------------------------
Thesis
- Title: Reading Faces using Hard Multi-Task Metric Learning for Kernel Regression
- Director: Mohamed Chetouani
- Advisor: Kevin Bailly
- Domain: Pattern Recognition, Computer Vision
----------------------------------------------------------------------------------------------------------------------------------------------------------
Research
- Automatic Action Unit Intensity Prediction (FG'15)
- Extending Metric Learning for Kernel Regression with a novel Multi-Task regularization
>>> The proposed system won the Facial Expression Recognition and Analysis challenge (FERA'15)
- Facial Landmark Localization (ICIP'13)
- New framework for an efficient use of LBP/LPQ with binary maps
- Triplet-based voting method for shape constraint inclusion
- Multimodal Prediction of Human Emotion in a Continuous Space (ICMI'12)
- Including dynamics with spectral features
- Learning temporal delay probability distributions of continuous labels
>>> The proposed system won the Audio/Visual Emotion Challenge (AVEC'12)
----------------------------------------------------------------------------------------------------------------------------------------------------------
Education
- 2013 - 2015 PhD in Pattern Recognition
- 2012 Training period (ISIR)
- 2008 - 2012 Engineering degree (Ecole Centrale de Lyon)
- 2012 Master's degree, Systems and Images (INSA Lyon)
- 2010 Training periods in signal processing (IRCAM, Télécom ParisTech)
- 2006 - 2008 Preparatory classes (Lycée Saint-Louis) Mathematics and Physics (MP*)