Title:
EXTRACTION OF EMOTIONAL FEATURES OF HUMAN FACE

Authors:
Yu He and Hongli Zhu

Abstract:
This paper introduces the face recognition dataset and the feature extraction techniques. and the deep learning method for face recognition will be introduced. The specific Pycharm software development platform is used, the system program is implemented by Keras, Open Cv, PyQt5 and other libraries, and the mini_XCEPITIOM convolutional neural network developed by Google is used to classify face emotions, The system can recognize the face by getting the picture from the camera in real time or by reading the input face, and the accuracy rate is about 65%

Keywords:
facial expression, convolutional neural network, deep learning, accuracy

DOI:
http://dx.doi.org/10.52267/IJASER.2021.2403