Human has various type of emotions by which his/ her mood can identify. There are three kinds of emotions such as basic, ambiguous, and complex type of emotions. Anger, joy, fear, sadness, neutral, surprise, and disgust are included in basic types of emotions. In the complex type of emotions pride, shame, anxiety, etc. are included. The ambiguous emotions are hope and compassion. The process which is used to detect human facial emotions is known as facial expression recognition. For a human, it is easy to understand the facial expression but for a machine it is difficult. The advancement in technology made expression recognition easier with the help of artificial intelligence. There are various applications of facial expression recognition systems such as user verification, mug shot matching, and user access control. There are several existing techniques of human facial expression recognition, but still, it is a challenging task. In the proposed work, an advanced learning-based CNN (Convolutional Neural Network) model with hybrid features is used to recognize human facial expressions. In images, some features are connection-oriented and some are edge-oriented, therefore hybrid methodology is used. The PCA (principal component analysis) and LBP (Local binary pattern) are used to extract the hybrid features from the image. For the classification of facial expression, a deep learning-based CNN model is used. Numerous fully connected layers are applied to the surface of one another in deep neural networks, for efficient recognition of complicated parts. The proposed model is evaluated on different parameters such as MSE, FAR, FRR, and accuracy.

Face expression recognition emotions hybrid features using LBP+PCA CNN (Convolutional Neural Network).