Emotion Detection Using Convolution Neural Network, Expert System and Deep Learning Approach

This work presents a facial expression identification system using the Facial Action Coding System with the use of the Bezier curves approximation method. This technique uses the features of the human face. These extracted face expressions are done with the idea of the face geometry and are also approximated by 3rd order Bezier curves by illustrating the relationship between the face feature movement and by observing the change of expressions. For face feature identification, color segmentation is done with the help of fuzzy logic classification which minimizes color similarities. Result outcomes define that this technique can identify the facial expressions with an accuracy of more than ninety cases. From human face structure, we divide into four regions such as right eye, left eye, nose, and mouth areas from the face image. Firstly, comes the face detection and the detection of the skin region. We crop the facial skin region and connect the largest skin region to detect the skin surface of the human face. When the emotion is perceived, the system recommends a play-list for the images. Based on the facial emotions, the Musical recommendation system creates a list of suggestions for music that are ranked from top to bottom.