Can I read DiM-Gesture: Speech-Driven Gesture Synthesis on EtoBox?
DiM-Gesture: Speech-Driven Gesture Synthesis by dangnguyen.blacksonia is a document available to read on EtoBox.
What is DiM-Gesture: Speech-Driven Gesture Synthesis about?
DiM-Gesture is a novel end-to-end generative model designed for personalized 3D gesture synthesis from raw speech audio, utilizing a Mamba-based fuzzy feature extractor and an Adaptive Layer Normalization (AdaLN) Mamba-2 diffusion architecture. This approach addresses challenges in gesture-speech synchronization and memory efficiency compared to existing Transformer-based methods, while maintaining high fidelity in gesture generation. Extensive evaluations demonstrate that DiM-Gesture outperforms state-of-t
- Author
- dangnguyen.blacksonia
- Language
- EN