Enregistré dans:
Détails bibliographiques
Auteurs principaux: Akhoundi, Elly, Ling, Hung Yu, Deshmukh, Anup Anand, Butepage, Judith
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2506.09075
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866913889081360384
author Akhoundi, Elly
Ling, Hung Yu
Deshmukh, Anup Anand
Butepage, Judith
author_facet Akhoundi, Elly
Ling, Hung Yu
Deshmukh, Anup Anand
Butepage, Judith
contents Motion in-betweening is a crucial tool for animators, enabling intricate control over pose-level details in each keyframe. Recent machine learning solutions for motion in-betweening rely on complex models, incorporating skeleton-aware architectures or requiring multiple modules and training steps. In this work, we introduce a simple yet effective Transformer-based framework, employing a single Transformer encoder to synthesize realistic motions for motion in-betweening tasks. We find that data modeling choices play a significant role in improving in-betweening performance. Among others, we show that increasing data volume can yield equivalent or improved motion transitions, that the choice of pose representation is vital for achieving high-quality results, and that incorporating velocity input features enhances animation performance. These findings challenge the assumption that model complexity is the primary determinant of animation quality and provide insights into a more data-centric approach to motion interpolation. Additional videos and supplementary material are available at https://silk-paper.github.io.
format Preprint
id arxiv_https___arxiv_org_abs_2506_09075
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SILK: Smooth InterpoLation frameworK for motion in-betweening A Simplified Computational Approach
Akhoundi, Elly
Ling, Hung Yu
Deshmukh, Anup Anand
Butepage, Judith
Graphics
Computer Vision and Pattern Recognition
Machine Learning
Motion in-betweening is a crucial tool for animators, enabling intricate control over pose-level details in each keyframe. Recent machine learning solutions for motion in-betweening rely on complex models, incorporating skeleton-aware architectures or requiring multiple modules and training steps. In this work, we introduce a simple yet effective Transformer-based framework, employing a single Transformer encoder to synthesize realistic motions for motion in-betweening tasks. We find that data modeling choices play a significant role in improving in-betweening performance. Among others, we show that increasing data volume can yield equivalent or improved motion transitions, that the choice of pose representation is vital for achieving high-quality results, and that incorporating velocity input features enhances animation performance. These findings challenge the assumption that model complexity is the primary determinant of animation quality and provide insights into a more data-centric approach to motion interpolation. Additional videos and supplementary material are available at https://silk-paper.github.io.
title SILK: Smooth InterpoLation frameworK for motion in-betweening A Simplified Computational Approach
topic Graphics
Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2506.09075