Apple’s machine learning researchers have developed a new application called Keyframer, which allows users to animate simple drawings by providing text descriptions. Tiffany Tseng, Ruijia Cheng, and Jeffrey Nichols from Apple have published a paper detailing the Keyframer app on the arXiv preprint server, showcasing the capabilities of the LLM GPT-4 in animation tasks.
Using this innovative app, users can input a basic drawing in SVG format along with a text prompt describing the desired animation. For example, a drawing of a rocket on a launch pad can be accompanied by a prompt like “Make the rocket launch into the sky with a bunch of fire blowing out beneath it.” Keyframer then generates the animation based on the provided instructions.
The researchers highlight that Keyframer leverages the LLM to imagine and execute the animation by generating CSS animation code. This allows for easy portability of the animation to different devices and enables users to manually edit the code to enhance or modify the animation. The app also supports iterative animation, enabling users to refine the project by adding new prompts for each iteration.
In their research paper, the team suggests that Keyframer has the potential to revolutionize the animation industry. If Apple makes the app available across all its hardware platforms, it could eliminate the need for multiple animation tools. Keyframer not only caters to professional animators but also empowers amateurs to create high-quality animated content effortlessly.
By combining the power of machine learning with intuitive user interfaces, Keyframer opens up new possibilities for creating captivating animations with minimal effort. This app could pave the way for a more accessible and efficient animation creation process, benefiting both experienced animators and beginners looking to delve into the world of animation.
*Note:
1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it