Article reprint source: Model Evolution

Source: CG World

How many people still remember Hirokazu Yokohara, the Japanese artist who is known as the almighty god for playing with models, animations, special effects and AI videos?

For those of you who have forgotten, let me remind you: I won’t say much about who the “uncle who turned into a sweet girl in a second” was that year, who was moved and then heartbroken.

This is a video generated by a master using ControlNet and TemporalKit Ebsynth on the StableDiffusion Web UI. The stability of this picture is so high that I feel that even if I am hurt by the "photo fraud", I still have to be strong and give the master a thumbs up.

Do you recall a little bit? In fact, CG World has shared many works of this great god in previous articles. After all, Japan is the country where the second dimension is dominant. In the hands of the great god, everything can be second dimension.

Recently, this all-rounder has come up with new tricks and heights, using Stable Diffusion to generate hand-painted style animations and special effects. Let’s take a look with the editor.

AI Animation Beta

Before producing the finished AI animation, the master also went through a period of testing. The master tried to convert the CG exported from Blender into an animation style using Stable Diffusion.

Generate simple animations with prompts

Although the presentation effect of generating animation only through prompts is not satisfactory, it is not bad to think that you can easily generate animation-like effects by specifying the movement of the timeline with simple prompts.

Using AI to create fictional animations

Looking at these animations spanning the styles of the 90s to the 2010s, do you always feel a sense of déjà vu? But in fact, they are not real cartoons, but fictionalized by AI. Isn’t it incredible, as if they have crossed a "parallel space"?

Finally, let’s take a look at the animation with fire effects created by the master using Stable Diffusion’s AnimateDiff~

Okay, that’s all for today’s sharing. Although there may still be many flaws at this stage, with the development of AI technology, the possibility of generating more complex animations will become higher and higher in the future.