Využití video generativní AI pro motion graphics

This bachelor thesis responds to the growing trend of generative AI tools in creative industries. Specifically, it deals with the use of video-to-video generative method for motion graphics. The work describes several ways to use this method. Two of them, style transfer and morph dynamic transition, are implemented in the workflow designed within the practical part. This workflow was made for ComfyUI and Stable Diffusion models. The most important findings of this procedure are suitably selected values for control of style and structure of the video. This control was based on the visual and motion reference created by the author. Another benefit is the principle of omitting several empty frames, creating space for the morph dynamic transition generated by artificial intelligence. The resulting animation is available at this link: https://youtube.com/shorts/1U1WVSc84Zs. The main benefits of the work are the instructions for the whole process and the designed workflow for ComfyUI. These can serve as a guide to the future creators of motion graphics and animation. The by-benefit is the defense of video-to-video creations from the perspective of Czech copyright, which currently does not protect AI generated work for its unoriginality and lack of creative input.