Real time facial expression recognition using RealSense camera and ANN

Abstract

Facial expression is an essential part in communication. It is a challenging task in computer vision as well as in pattern recognition. Facial expression recognition has applications in many fields such as HCI, video games, virtual reality, and analyzing customer satisfaction etc. Proposed system focuses on emotion recognition using facial expressions which are captured by using Intel's RealSense SR300 camera. This camera detects landmarks on the depth image of a face automatically using Software Development Kit (SDK) of RealSense camera. Geometric feature based approach is used for feature extraction. The distance between landmarks is used as features and for selecting an optimal set of features brute force method is used. Proposed system is used Multilayer Perceptron (MLP) neural network algorithm using backpropagation method for classification. The experimental dataset is captured using RealSense SR300 camera. The proposed system recognizes three facial expressions namely neutral, happy, and surprised. The recognition rate achieved is 93.33%.