Rethinking Facial Expression Recognition in the Era of Multimodal Large Language Models: Benchmark, Datasets, and Beyond

The recent advancements in Multimodal Large Language Models (MLLMs) are reshaping the landscape of facial expression recognition (FER) by integrating it with computer vision and affective computing. This shift towards unified approaches, particularly through the transformation of traditional FER datasets into visual question-answering formats, opens up exciting possibilities for more effective and comprehensive understanding of human emotions. This matters because it not only enhances the accuracy of emotion detection but also broadens the applications of FER in various fields, from security to mental health.