Cylab Biometrics Center, Carnegie Mellon University, USA Sep 2016 - 2018
- Deep Learning Models for Face Matching
- Soft biometrics with age and gender estimation
- Tractable Deep Generative Models for Longitudinal Face Modeling
- Generative Adversarial Networks for Face Modeling
- Deep Learning Models for Scene Parsing
Concordia University, Canada Jan 2013 - 2016
- Deep Generative Model for Longitudinal Face Modeling
- Deep Boltzmann Machines for Face Modeling: proposed a novel Deep Appearance Models (DAMs) approach to accurately capture both shape and texture of face images under large variations. This model can be used as an efficient replacement for Active Appearance Models (AAMs).
- Low rank matrix factorization, Compressed sensing, Sparse representation and dictionary learning: evaluating the sparse representation based system for face recognition with occlusions.
- Handwritten character recognition: applied the theories of sparse representation and dictionary learning to handwritten character recognition.
National Institute of Informatics, Tokyo Japan Feb 2012 - Aug 2012
- Eye Localization in Video: proposed to combine both tracker and detector for robust eye localization in video. The new system can overcome challenging factors such as: extreme head pose change, closed eyes and occlusion.
- Video Aesthetic Assessment: proposed a system to automatically extract the beautiful scene in broadcast videos. Four scene-level features for scene aesthetic assessment including (1) length of scene, (2) scene difference, (3) overall picture of scene and (4) object interestingness are designed. These four features will fully exploit the temporal information and provide a better understanding of human impression from scene.
- Face Recognition
FPT Software, Ho Chi Minh City, Vietnam Jun 2007 - Aug 2007
Software and Web Design Engineer Intern
- Designed and developed a web-based translated document management system.
University of Science, Vietnam
Research Assistant Sep 2009 - Feb 2012
- Facial Age Estimation: proposed an advanced age-estimation approach that combines global and local features derived from a facial image.
- Digital Image Tampering Detection using Color Filter Array.
Undergraduate Student Sep 2004 - Aug 2008
- Video Watermarking: applied DCT and DWT techniques to embed copyright information to video and proposed a new method that combines these techniques to improve the scheme’s robustness. • A project in Handwritten characters recognition using Feed-Forward Neural Network. • A course project in Image Processing: develop Traffic Sign Detection with four kinds of sign: circle, triangle, rectangle and square.
Cylab Biometrics Center, Carnegie Mellon University, USA
- Face Matching SDK
- Face Analysis SDK for age and gender estimation
- Long range face matching
Vietnam National University, Vietnam
- Efficient Face Retrieval In Large Scale Video Archives (2012 - 2013)