Emotion Detection on live video using Deep Learning

Abstract

In modern days, feeling exposure is a ground of curiosity and is used in fields such as cross-examining prisoners and teenagers observing human-computer relations. The anticipated work designates the exposure of mortal sentiments from an instantaneous video or stationary video with the help of a convolution neural network (CNN) and haar cascade algorithm. The foremost part of the announcement constitutes field appearance. The suggested work aims to categorize a given video or a live video into one of the emotions (natural, angry, happy, fearful, disgusted, sad, surprise). Our work also distinguishes multiple faces from live video and organize their emotions. Our recommended work also imprisonments the pictures from the video every second, hoard them into a file, and generates a video from those pictures along with their respective.