Emotion Recognition Using CNN

Abstract

Humans have the capacity to produce a varied range of facial actions during communication that vary in intensity, complexity and meaning. Machines can be trained to understand human emotions by using processed images converted into data as an input to a neural network. The system will be using cameras to acquire images. The images will be further processed and filtered to prepare the images for data extraction. This data is then provided to the neural network for training and testing purpose. After training our system to recognise emotions, it will be implemented to increase security by recognizing mass fear and notifying the authorities.