SCPT

This repository provides the official implementation of our method proposed in the paper: "Adaptive Physical-Facial Representation Fusion via Subject-Invariant Cross-Modal Prompt Tuning for Video-based Emotion Recognition"