Implementation of Real-Time Facial Emotions for Feedback System using Deep Learning

Abstract

Recognition of Sentiment or Emotion from Human Facial Expression has numerous applications in fields, such as Security, Human Computer Integration(HCI), Advertisement Strategies, Accessibility, Medical Diagnosis, etc. Written feedback for the Lectures given in the Classroom is not efficient to know whether the lecture had reached the desired audience. In this paper, the feedback applied with deep learning techniques that includes Convolution Neural Network (CNN )helps the lecturer to analyse  process detect and classify the human’s key emotions such as Happy, Sad, Anger and Neutral. The system is designed to capture and detect the faces and their emotions to determine the satisfaction of the listener. Object Detection using Haar feature-based cascade classifiers is an effective object detection method which is used to detect the faces from the Image. Using the Deep Learning Model, Keras, a high-level Neural Network runs on the top of Tensor Flow framework. Emotion of facial features is also determined. From the collective data, the result clearly exhibits  the feedback of the lecture were effective.

Keywords: Human Computer Integration, Convolution Neural Network, deep learning.