2022-11-02-Video Understanding
Video Understanding method
- single frame
- early fusion
- late fusion
- 3D CNN
- C3D (The VGG of 3D CNNs)
- Two-Stream Networks
- I3D 膨胀卷积
- Visio transformer for video
GAN 生成式模型
VAE 自动编码器(Variational Autoencoder)
Pixel RNN, Pixel CNN
Generative Adversarial Networks
一个生成器、一个监督器、互相对抗学习