Motion Style Slider: Endpoint-Supervised Continuous Style Control for Human Motion Diffusion

2026-09-25

AI Summary

This study addresses the challenge of fine-grained, continuous control over style intensity in human motion diffusion models by proposing an endpoint-supervised motion style transfer framework. By integrating latent-space style direction construction, diffusion-based denoising, and latent intensity regularization, the method achieves smooth and monotonic style modulation using only scalar intensity values, eliminating the need for intermediate ground-truth annotations. Furthermore, the framework supports training on heterogeneous datasets and enables out-of-range extrapolation. Experimental results demonstrate that the proposed approach significantly enhances both the controllability and realism of generated motions while effectively preserving content consistency across interpolation and extrapolation tasks.

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.

Problem

• human motion diffusion
• style transfer
• continuous style control
• style intensity
• motion generation

Innovation

• Motion Style Slider
• Endpoint-Supervised Control
• Human Motion Diffusion
• Style Transfer
• Latent Intensity Regularization