The field of facial expression analysis is witnessing a significant shift towards more comprehensive and innovative approaches. Researchers are moving beyond traditional methods, such as the Facial Action Coding System (FACS), and exploring alternative coding systems that can capture localized and interpretable facial movements. This has led to the development of data-driven coding systems that can reconstruct all observable facial movements, overcoming the limitations of automated FACS coding. Furthermore, the use of deep learning architectures and attention mechanisms has improved the accuracy of micro-expression recognition and facial behavior analysis. The emphasis is on creating more trustworthy and transparent systems, with a focus on fairness, explainability, and safety. Noteworthy papers in this area include:

• The introduction of the Facial Basis, a data-driven coding system that outperforms traditional AU detectors in predicting autism diagnosis.
• FaceSleuth, a dual-stream architecture that enhances motion along the vertical axis and delivers state-of-the-art performance in micro-expression recognition.
• OpenFace 3.0, a lightweight multitask system for comprehensive facial behavior analysis that exhibits improvements in prediction performance and inference speed.