Abstract
With advances in artificial intelligence and digital media, dance creation increasingly intersects with computational methods for motion and emotion modeling. Existing approaches often emphasize kinematic accuracy but overlook emotional integration. This study presents an AI-driven framework that combines motion capture–based feature extraction with psychological emotion modeling to generate emotion-aware dance movements. Experiments on public datasets show improved emotional consistency and expressive diversity, demonstrating that affective computing effectively guides human–AI co-creation for digital performance and interactive media applications.