Subject Identification Across Large Expression Variations Using 3D Facial Landmarks

In this work, we propose to use 3D facial landmarks for the task of subject identification, over a range expressed emotion. Landmarks are detected, using Temporal Deformable Shape Model and used train Support Vector Machine (SVM), Random Forest (RF), Long Short-term Memory (LSTM) neural network identification. As interested in identification with large variations expression, conducted experiments on 3 emotion-based databases, namely BU-4DFE, BP4D, BP4D+ 3D/4D face databases. We show that our proposed method outperforms current state art methods BU-4DFE BP4D. To best knowledge, is first work investigate BP4D+, resulting baseline community.