A Survey on Methodologies and Difficulties in Facial Emotion Recognition

In Human-Computer Interaction (HCI), Facial Emotion Recognition or Facial Expression Recognition (FER) is a critical area within Pattern Recognition. This paper explores the importance of recognizing facial expressions, traditionally observed through eyes, eyebrows, and mouth movement. Although face-to-face analysis is effective, machine learning (ML), particularly Convolutional Neural Network (CNN) based approaches, has achieved success in this field. Despite this progress, several difficulties persist in the recognition process, including occlusion, tilt, illumination, inter-class variance, and more. This study systematically reviews existing methodologies, identifying both a basic structure for recognition and common challenges across various datasets. Additionally, it compares diverse techniques, offering insights into their efficacy in solving frequent difficulties. The paper's findings contribute to the understanding of facial expression recognition and can guide future research in developing more robust and effective solutions.