Gan Generating Video Motion Driven Still Image To Video

This article delves into various methods and tools, including CoNR, digan, MoCoGAN, TRPG Replay Generator, Montage.ai, and TikTok montages, that employ techniques such as GANs and thin-plate spline motion models to create videos from images or portraits.

---

thin plate spline motion model

video animation generation only using few character portraits: 根据角色设定图画出人物动画（有动画驱动器）

https://github.com/megvii-research/CoNR

galgame video generator using pygame（自动化类galgame动画生成器）:

https://github.com/w4123/TRPG-Replay-Generator

digan, could generate taichi videos

suggest you to segment video first and then use this to do the freaking work.

https://github.com/sihyun-yu/digan?ref=pythonawesome.com

https://sihyun-yu.github.io/digan/

https://pythonawesome.com/official-pytorch-implementation-of-generating-videos-with-dynamics-aware-implicit-generative-adversarial-networks/

MoCoGAN can generate the same object performing different actions, as well as the same action performed by different objects:

https://github.com/sergeytulyakov/mocogan

still image to talking with hands moving video generation compared to FOMM, can even animate non-human objects:

https://snap-research.github.io/articulated-animation/

montage.ai generate video by music with deep analyzer:

https://github.com/Tartar-san/montage.ai

tiktok hashtag montage:

https://github.com/andreabenedetti/tiktok-montage