← ArchivePaper2024
Muscles in Time: Learning to Understand Human Motion by Simulating Muscle Activations
David Schneider, Simon Reiss, Marco Kugler, Alexander Jaus, Kunyu Peng, Susanne Sutschet, M. Saquib Sarfraz, Sven Matthiesen, Rainer Stiefelhagen
MinT: large-scale synthetic dataset of muscle activation sequences enriched from motion capture via OpenSim simulation, covering 227 subjects and 402 muscle strands.
Abstract
Exploring the intricate dynamics between muscular and skeletal structures is pivotal for understanding human motion. This domain presents substantial challenges, primarily attributed to the intensive resources required for acquiring ground truth muscle activation data, resulting in a scarcity of datasets. In this work, we address this issue by establishing Muscles in Time (MinT), a large-scale synthetic muscle activation dataset. For the creation of MinT, we enriched existing motion capture datasets by incorporating muscle activation simulations derived from biomechanical human body models using the OpenSim platform, a common approach in biomechanics and human motion research. Starting from simple pose sequences, our pipeline enables us to extract detailed information about the timing of muscle activations within the human musculoskeletal system. Muscles in Time contains over nine hours of simulation data covering 227 subjects and 402 simulated muscle strands. We demonstrate the utility of this dataset by presenting results on neural network-based muscle activation estimation from human pose sequences with two different sequence-to-sequence architectures. Data and code are provided under https://simplexsigil.github.io/mint.
How to read this
- Category
- Dataset: a synthetic muscle-activation dataset
- Contributions
- Muscles in Time (MinT), a large-scale synthetic muscle-activation dataset built by enriching motion-capture with OpenSim simulations
- Over nine hours of simulation data covering 227 subjects and 402 simulated muscle strands
- Baseline muscle-activation estimation from pose sequences using two seq-to-seq architectures, with data and code released
- Context
- Addresses the scarcity of ground-truth muscle data using biomechanical body models, relating to musculature-driven motion work such as Ryu et al.'s functionality-driven musculature retargeting.Builds on: Functionality-Driven Musculature Retargeting
- Correctness
- Activations are simulated via OpenSim rather than measured, so the data inherits the biomechanical model's assumptions; a reader should treat it as a synthetic benchmark whose realism depends on the underlying musculoskeletal model.
- Clarity
- Accessible as a dataset paper; a first pass conveys scope and pipeline, a second pass for the simulation setup and baseline architectures.
- How to read it
- Read the dataset construction (OpenSim enrichment, coverage) and the baseline task definition first; a deep methods pass is only needed if you plan to train on or extend it.
Builds on
Built upon by
Nothing yet.
Related work
- AddBiomechanics: Automating model scaling, inverse kinematics, and inverse dynamics from human motion data through sequential optimization 2023 / PLOS ONE
- From Skin to Skeleton: Towards Biomechanically Accurate 3D Digital Humans 2023 / SIGGRAPH Asia
- TEMOS: Generating Diverse Human Motions from Textual Descriptions 2022 / CVPR
- MoConVQ: Unified Physics-Based Motion Control via Scalable Discrete Representations 2024 / TOG
Keywords
This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →