Deep Learning-Based AI Model for Recognizing Emotional Changes via Micro-expressions and Micro-actions

Accurate recognition of emotional changes has become an urgent and essential need in multiple practical fields and interdisciplinary domains.Traditional emotion recognition primarily relies on monitoring explicit expressions and the subjective statements of subjects, and subjective biases often significantly influence its overall accuracy and reliability.With the powerful help of modern deep learning technology, the ability of multi-layer neural networks to automatically extract complex and subtle features enables the detailed observation of the subjects' micro-expressions and micro-actions.These two types of extremely subtle behavioural cues are generally not controlled by subjective consciousness and can therefore be relatively objectively reflective of the real psychological and emotional state of the subjects.This study aims to construct an AI model based on deep learning, integrating the comprehensive multimodal information of micro-expressions and micro-actions to more accurately and robustly recognise emotional changes and further assist in improving, validating, and correcting test results in real-world applications.