Three-Dimensional Emoji Prediction Using Facial Emotion Recognition

Abstract

Reliable communication relies on thoughts of human while communicating with a group of people. Facebook, Instagram, and Twitter are some of the social media where various types of human emotions are being shared. Two-dimensional emojis are widely used in communication rather than sharing human faces on social media. Sharing twodimensional emojis is more comfortable without sharing live human facial emotions using a video stream. As people do not share their own real-time videos on social media, it is more realistic if people can share three-dimensional emojis based on real-time facial emotions on social media. Real-time facial emotion recognition using facial expressions helps to embed three-dimensional emojis with relevant facial expression type. This thesis investigates human facial emotion recognition using facial expressions and integrating real-time facial emotions with three-dimensional emojis. Modules of the research design consists of face detection module, emotion classification module and three-dimensional emoji prediction module. Face detection module detects the face of a human using a real-time video stream. Emotion classification module uses the detected human face to extract the human facial expression with facial emotions. Three-dimensional emoji prediction module consumes extracted facial emotion type to predict associated three-dimensional emoji to the user. Evaluation of the research study was conducted using a sample group. The response given by the sample group was used to analyze the research with results. Evaluation process was performed via a selected set of parameters gained by the research methodologies. The accuracy level and effectiveness level of the emotional change on 3D emoji were calculated to evaluate the research study to maintain a better relationship with the input of the research study and the expected output. This research focused on three-dimensional emoji prediction using facial emotion recognition. Research study emphasized on improving human machine interaction between social communication media.