About this document
Dynamic Future Net for Human Motion Generation by ryl0903 is a document available to read on EtoBox.
The document presents a new deep learning model called Dynamic Future Net (DFN) for generating diversified human motion. DFN explicitly models both short-term and long-term stochasticity in human motion dynamics. It learns latent representations of poses and dynamics and their conditional distributions. DFN can generate high-quality motions with different styles and arbitrary durations from limited training data.
- Author
- ryl0903
- Language
- EN