Enhancing modern day machines or computers to recognize various facial expressions and to understand human emotions from them in real time is an exigent research subject. Through this paper, I put forward a solution to recognize emotions by understanding different facial expressions by collecting live video through a Flask App created. I deploy a Flask App to video stream live feed captured through the local camera attached to the machine or computer system. The video captured is fed to various image extraction techniques. The facial features are identified by different operations provided by OpenCV and the region consisting of parts of the face are made to surround or enclose by a contour. This region, enclosed by the contour is used as an input to Convolutional Neural Network (CNN). The CNN model created consists of six activation layers, of which four are convolution layers and two are fully controlled layers. Each layer is designed to undergo several training techniques. The main objective of this project is to demonstrate the accuracy of Convolutional Neural Network model designed. The paper is concluded by discussing the outcomes of our project and the ways to improve the efficiency of the model. The scope of this project is also analyzed to enhance technologies developed in the near future.