Abstract
The implementation of artificial intelligence (AI) technology in the field of animation-making has resulted in the development of innovative tools such as Deforum and Mixamo that potentially offer new possibilities and improve the efficiency of animators’ workflows. EbSynth (EbS) is another one of these tools that allows users to animate existing footage in the rotoscoped animation technique using just a few styled keyframes. While EbS is not generally classified as an AI application, it utilizes Example-based Synthesis algorithms that can be considered AI-informed according to the broadest definition of the term. Our research goal centers on the use of enhanced trackable patterns and shapes in EBS and their impact on the efficiency and quality of rotoscoped animation. We seek to identify the most effective patterns and shapes for this process while establishing workflow guidelines for EbS users. Adopting a practice-led research approach, we employ our creative practice to generate insights into the effectiveness of trackable patterns and shapes applied to rotoscoped animation using EbS. Our study encompasses male and female models performing various actions, including facial muscle movements and emotions, with a focus on patterns, trackable markers, contours, and character design shapes. Through a series of experiments and iterative analyses, we evaluate the impact of enhanced trackable patterns and shapes on the quality and efficiency of rotoscoped animation. Our findings support the hypothesis that this approach improves the rotoscoping process, offering valuable insights for artists and animators.
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
KEYWORDS
Animation, Artificial Intelligence, EbSynth, Example-based Synthesis, Rotoscopy
