Out of curiosity, I've been informally exploring different ai tools to see if they can take an illustration (either non-ai generated or ai-generated) and bring it to life as an animation. I'm looking at the output quality, prompts, workflow techniques, models, style-tunings, LoRAs. This exploration was done in Krea using variety of tools, and the Pika 2.2 model.
Source image created in Krea.
Here's a couple of other exploration examples for comparison.
This one used the Wan 2.1 model in Krea to generate the video from an illustration image.
This one used the Hailuo model in Krea to generate the video from an illustration image.
This one was created without using a seed image. Generated from a prompt in Krea. It used the Pika 2.2 model.
Once I settle on the visual style and am getting consistency, then I'll probably create 5-second clips, and each new clip will start with the last frame of the previous video. This way, you can stitch them together in a video editing tool.