Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AStF: Motion Style Transfer via Adaptive Statistics Fusor

Hanmo Chen, Chenghao Xu, Jiexi Yan, Cheng Deng

TL;DR

The paper introduces AStF, a motion style transfer framework that enriches statistical modeling beyond mean and variance by incorporating skewness and kurtosis alongside traditional statistics. It combines a Style Disentanglement Module (SDM) with a High-Order Multi-Statistics Attention (HOS-Attn) to fuse spatiotemporal statistics from style and content motions, and employs a Motion Consistency Regularization (MCR) discriminator to enforce style coherence. Extensive experiments on Xia and BFA datasets show superior style fidelity and content retention compared to state-of-the-art methods, supported by ablations that highlight the contributions of SDM, high-order statistics, and the MCR loss. The work advances motion stylization by explicitly modeling higher-order temporal statistics and their integration into content motion, with practical impact for realistic character animation and related applications.

Abstract

Human motion style transfer allows characters to appear less rigidity and more realism with specific style. Traditional arbitrary image style transfer typically process mean and variance which is proved effective. Meanwhile, similar methods have been adapted for motion style transfer. However, due to the fundamental differences between images and motion, relying on mean and variance is insufficient to fully capture the complex dynamic patterns and spatiotemporal coherence properties of motion data. Building upon this, our key insight is to bring two more coefficient, skewness and kurtosis, into the analysis of motion style. Specifically, we propose a novel Adaptive Statistics Fusor (AStF) which consists of Style Disentanglement Module (SDM) and High-Order Multi-Statistics Attention (HOS-Attn). We trained our AStF in conjunction with a Motion Consistency Regularization (MCR) discriminator. Experimental results show that, by providing a more comprehensive model of the spatiotemporal statistical patterns inherent in dynamic styles, our proposed AStF shows proficiency superiority in motion style transfers over state-of-the-arts. Our code and model are available at https://github.com/CHMimilanlan/AStF.

AStF: Motion Style Transfer via Adaptive Statistics Fusor

TL;DR

The paper introduces AStF, a motion style transfer framework that enriches statistical modeling beyond mean and variance by incorporating skewness and kurtosis alongside traditional statistics. It combines a Style Disentanglement Module (SDM) with a High-Order Multi-Statistics Attention (HOS-Attn) to fuse spatiotemporal statistics from style and content motions, and employs a Motion Consistency Regularization (MCR) discriminator to enforce style coherence. Extensive experiments on Xia and BFA datasets show superior style fidelity and content retention compared to state-of-the-art methods, supported by ablations that highlight the contributions of SDM, high-order statistics, and the MCR loss. The work advances motion stylization by explicitly modeling higher-order temporal statistics and their integration into content motion, with practical impact for realistic character animation and related applications.

Abstract

Human motion style transfer allows characters to appear less rigidity and more realism with specific style. Traditional arbitrary image style transfer typically process mean and variance which is proved effective. Meanwhile, similar methods have been adapted for motion style transfer. However, due to the fundamental differences between images and motion, relying on mean and variance is insufficient to fully capture the complex dynamic patterns and spatiotemporal coherence properties of motion data. Building upon this, our key insight is to bring two more coefficient, skewness and kurtosis, into the analysis of motion style. Specifically, we propose a novel Adaptive Statistics Fusor (AStF) which consists of Style Disentanglement Module (SDM) and High-Order Multi-Statistics Attention (HOS-Attn). We trained our AStF in conjunction with a Motion Consistency Regularization (MCR) discriminator. Experimental results show that, by providing a more comprehensive model of the spatiotemporal statistical patterns inherent in dynamic styles, our proposed AStF shows proficiency superiority in motion style transfers over state-of-the-arts. Our code and model are available at https://github.com/CHMimilanlan/AStF.

Paper Structure

This paper contains 16 sections, 11 equations, 2 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Image Style Transfer
  4. Motion Style Transfer
  5. Method
  6. Preliminary
  7. Method Overview
  8. Style Disentanglement Module
  9. High-Order Multi-Statistics-Attention
  10. Motion Consistency Regularization
  11. Loss Function
  12. Experiments
  13. Experimental Settings
  14. Evaluation Results
  15. Ablation Study
  16. ...and 1 more sections

Figures (2)

  • Figure 1: Statistics distribution differences across styles by t-SNE on BFA dataset aberman2020unpaired.
  • Figure 2: Method Overview. (a) Overall framework of AStF mainly comprising Content and Style Statistics-Encoder, Style Disentanglement Module (SDM), High-Order Multi-Statistics-Attention(HOS-Attn) and decoder. (b) Detailed structure of SDM and Simple SDM. (c) Detailed structure of HOS-Attn. (d) Pipeline of Motion Consistency Regularization (MCR) discriminator.