Facial micro-expression analysis

Abstract

Facial expressions offer valuable insights into an individual’s emotional state. However, emotions are often complex and may be deliberately masked or unconsciously suppressed, making it difficult to interpret a person’s true feelings. This challenge is particularly evident in medical and psychological contexts, where patients might hesitate to fully express their emotions—especially during remote consultations. Such reluctance can hinder accurate diagnosis and appropriate care. Facial micro-expression analysis presents a promising solution by capturing brief, involuntary facial movements that reflect genuine emotions, even when individuals attempt to conceal them. Recent advancements in technology have enabled the integration of facial expression analysis into psychological care, providing therapists with innovative tools to augment traditional diagnostic methods. Micro-expression recognition has emerged as a particularly powerful technique, capable of revealing subtle emotional cues that may be overlooked during therapy sessions. Recognizing the clinical value of this approach, this project introduces a real-time iOS mobile application designed to analyze micro-expressions and provide therapists with deeper emotional insights into their patients. The application leverages trained Convolutional Neural Networks (CNNs) to detect and classify micro-expressions from facial video data captured on Apple devices. By offering therapists objective and immediate emotional feedback, the system aims to enhance diagnostic accuracy and improve the overall quality of psychological care.