Keywords: Deep learning, Animation, Latest Method, python
This is the ReadingNotes from video games105.
We have learned several animation method:
- interpolation animation(keyframes, in-betweens)
- kinematics animation(FK, IK)
- data-based animation(motion capture)
- physical-based animation
this chapter focus on:
- learning based animation(Deep learning & Reinforcement learning)
State Machine, Motion Matching is using the existing mocap motion(mocap from professional actors), but the mocap motion is limited, we want to generate new natural motions. So the problems becomes how to generate new natrual motions based on the motions we already have?
PCA
Gaussian Mixture Model
Autogressive Model
Neural Network
VAE
GAN
Normalizing Flow
Diffusion Model 扩散模型
