Human Facial Emotion Recognition using Deep Neural Networks

Abstract

Humans experience a plethora of different, confusing and nearly indiscernible emotions. Interpreting and understanding these emotions is a rewarding challenge as they play an important role in developing and maintaining interpersonal relationships. Hence, extracting and understanding emotions is paramount to the interaction and communication between human and machine. This understanding aids applications such as brand humanization, sales, advertising and marketing by helping to gauge the response of existing or prospective clients, medical industry, e-learning, law enforcement, automatic counseling system, drunk-driving detection, pain or stress detection, brand value analysis from consumer reactions and the entertainment industry, social interactions and facilitate rational decision making and perception. The recognition of facial emotion is quite difficult problem since there will be a variation of same emotion of single person owing to occlusion, illumination, aging, pose, gender etc. This paper proposes a deep learning model to detect the seven basic emotion classes. The dataset used are CK+, JAFFE and it achieves higher accuracy when compared to the existing approaches.