Recognition Of Facial Expressions Using A Deep Neural Network

Facial expression recognition technology has boomed over the past few years because of human-computer engagement. Computer vision advancements have made it possible that machines can now understand the human’s actions., expressions, etc. Research in this area is also a hot topic because it offers a wide range of applications and shows that CNN provides impressive results compared to traditional methods. So keeping it as a motivation, in our work, we aimed for such Deep CNN architecture, which can work on real-world images like images having various resolution, angles, poses, illumination, and brightness, etc. So for this, we have implemented our CNN architecture with the Kaggle challenge presented dataset FER-2013 and trained the model to recognize the basic seven expressions. The proposed approach seems to be effective since we were able to achieve a validation accuracy of 70.15%. This approach not only can be applied to other datasets but also in real-world applications.