CNN Based Facial Expression Recognition System Using Deep Learning Approach

Abstract

In this fast-growing world of 21st century Machine Learning (ML) is one of the most powerful tools for solving a variety of real-life problems. Machine Learning helps both normal as well as differently-abled people. Facial Expression Recognition (FER) system can be widely used in different variety of research areas, such as diagnosis of mental disease and human physiological interaction detection. Machine Learning can automate processes by learning from different datasets. ML is commonly used for tasks like prediction, detection, regression and classification. The core idea of ML is to learn from previous and produce results accordingly. This result may be the output of new input or predictions based on past actions. With this, Deep Learning a subset of ML made possible to learn human facial expression from an image dataset and identify the facial expression performed in real life. The advancement in facial recognition systems helps in the purpose of diagnosis and also market feedback. But, due to various complexities and noise factors in facial expression images, there is still a need for improvement in this field. With different technological advancements in hardware and software, Facial Expression Recognition systems have been developed to help and support real-world applications.