A Python implementation for the method described in the paper:
Performance Comparisons of Facial Expression Recognition in Jaffe Database. https://www.researchgate.net/publication/220359507_Performance_Comparisons_of_Facial_Expression_Recognition_in_Jaffe_Database
- preprocessing
- Feature extraction
- Expression classification (SVM)
Download the database from the following link: http://www.kasrl.org/jaffe.html The database contains ten Japanese females. Seven different facial expressions (neutral, happy, angry, disgust, fear, sad, surprise)
This will include:
- Eyes detection
- rotate the face by the angle between eyes-axis and x-axis
- crop the face (as descriped in the paper)
- resize the copped face to (96,128)
- Equalize Hist
pip install PyWavelets
pip install dlib
pip install opencv-python
pip install imutils
in addition to numpy, sklearn and matplotlib