INSIGHT ON HUMAN ACTIVITIES USING DEEP LEARNING APPROACHES

Abstract

Humans consistently have the inborn capacity to 
perceive and recognize faces and emotions. 
Computers can now accomplish the same thing, which 
creates lots of new challenges in daily life. For 
example, emotion recognition can increase security, 
enable financial transactions without the need for real 
cards, and enable the identification of criminals and 
specialized treatment, among other things. Therefore, 
face detection and emotion recognition is a prominent 
and futuristic research topic. In the near future, open
source projects will get more importance rather than 
licensed ones. In this connection, it is proposed to 
utilize a python library for face detection and 
recognition. Kaggle web resources are providing 
open-source Face Emotion Recognition (FER) 
datasets. With the help of datasets, there are seven 
types of emotions: happy, sad, fear, disgust, angry, 
neutral, and surprise. It is proposed to use image 
augmentation to improve emotion recognition by 
building a six-layered Convolution Neural Network 
(CNN) in Python using the Keras toolkit.