Multimodal Functional Maximum Correlation for Emotion Recognition
Researchers have developed a new self-supervised learning framework called Multimodal Functional Maximum Correlation (MFMC) to improve emotion recognition from physiological signals. MFMC is designed to capture higher-order interactions across multiple modalities, unlike previous methods that focused on pairwise alignments. Experiments on public benchmarks show MFMC achieves state-of-the-art or competitive results, significantly improving accuracy in subject-dependent and subject-independent evaluations. AI
IMPACT This new framework could lead to more accurate and robust emotion recognition systems, impacting fields like mental health monitoring and human-computer interaction.