Personalized Emotion Recognition Utilizing Speech Signal and Linguistic Clues

Abstract

In this project, non-intrusive videos are taken in real-time, which detect the emotional state of an individual by analyzing the facial expression. We present the Convolution Neural Network (CNN), a machine learning algorithm used for automatic image classification. This supervised classification technique analyses and trains the classifier on the labeled images and extracts the features of the classifier. By using the learned information of the training, the newly provided image will be classified based on the features observed in the image. The prototype which we are going to develop detects whether a person is happy or sad or angry or surprise.