Summary
Changing how a character moves in animation to show different styles can be tricky, especially when you want smooth control over how strong the style looks. The researchers designed a system called Motion Style Slider that lets users control the style of motion continuously, like adjusting a volume knob, even though there aren't ready examples for every style level. They created a way to teach a computer to understand style directions and how much to apply them, making animations that can smoothly shift between styles without losing the original movement. This method works with existing motion generation tools and can handle mixed datasets and new extreme styles.
What this means in practice
- •For animation studios: Adjust animation character movements with precise, smooth control of style intensity for creative direction without needing new training data for every style level.
- •For game development teams: Create more nuanced and continuous stylistic effects in character motions that can be tuned in real-time, enhancing gameplay realism and expressiveness.
Authors
Chen-Chieh Liao, Yichen Peng, Yiyi Cai, Yûi Ono, Hiroki Hanaoka, Erwin Wu, Hideki Koike, Shuichi Kurabayashi
Abstract
Existing human motion diffusion methods provide strong motion generation quality, and recent style transfer models can inject target style cues, but fine-grained continuous control of style intensity remains underexplored. In production, style intensity is subjective across artists and directors, so the practical requirement is not a universal absolute unit, but a reliable monotonic control axis. We propose Motion Style Slider, a motion-to-motion style transfer framework for endpoint-supervised continuous control. Given a content motion and a style motion, we construct a style direction in a learned motion-style embedding space and condition diffusion generation with a scalar intensity. The training objective combines diffusion denoising with latent intensity regularization to encourage smooth and monotonic style scaling without requiring intermediate-intensity ground-truth motions. Our framework is compatible with pretrained motion diffusion backbones and supports heterogeneous style datasets, including the multi-actor style motion dataset. To test out-of-range usability, we additionally introduce a small real-capture over-reaction extension and evaluate large-intensity behavior against these unseen targets. Experiments measure controllability, interpolation/extrapolation behavior, content preservation, and motion realism, with ablations on direction construction and loss design.