LIMMT: Less is More for Motion Tracking

Hugging Face Daily Papers Papers

Summary

This paper introduces LIMMT, a data-centric study showing that training with high-quality, minimal subsets of motion data (under 3% of AMASS) outperforms using the full dataset for physics-based humanoid motion tracking, defining motion data quality through physics feasibility, diversity, and complexity.

We argue that high-quality motion data can steer tracking policies toward better optimization trajectories early in training. In this work, we introduce LIMMT (Less Is More for Motion Tracking). To our knowledge, this is the first data-centric study for physics-based humanoid motion tracking. We go beyond simply removing low-quality and erroneous clips, but define motion data quality through three dimensions: physics feasibility, diversity, and complexity. We show that even training with under 3% of AMASS yields better tracking performance than training with the full dataset. We further conduct data cleaning on the estimated web-sourced mocap data. Extensive experiments and analyses validate the effectiveness of our framework.
Original Article
View Cached Full Text

Cached at: 06/08/26, 07:14 AM

Paper page - LIMMT: Less is More for Motion Tracking

Source: https://huggingface.co/papers/2606.06953

Abstract

Training with high-quality motion data improves tracking policy optimization trajectories, with minimal data subsets outperforming full datasets in physics-based humanoid motion tracking.

We argue that high-quality motion data can steer tracking policies toward better optimization trajectories early in training. In this work, we introduce LIMMT (Less Is More forMotion Tracking). To our knowledge, this is the firstdata-centric studyforphysics-based humanoid motion tracking. We go beyond simply removing low-quality and erroneous clips, but definemotion data qualitythrough three dimensions:physics feasibility,diversity, andcomplexity. We show that even training with under 3% ofAMASSyields better tracking performance than training with the full dataset. We further conductdata cleaningon the estimated web-sourcedmocap data. Extensive experiments and analyses validate the effectiveness of our framework.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2606\.06953

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.06953 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.06953 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.06953 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

HumanTracker: Towards Comprehensive and Human-Aligned Motion Tracking Benchmark

Hugging Face Daily Papers

HumanTracker introduces a large-scale benchmark and a preference-aligned metric called HumanScore to evaluate humanoid motion tracking, focusing on perceptual quality and physical contact stability to better predict human preferences and identify failures missed by traditional kinematic metrics.

AnyMo: Geometry-Aware Setup-Agnostic Modeling of Human Motion in the Wild

Hugging Face Daily Papers

AnyMo is a geometry-aware framework for setup-agnostic human motion modeling using physics-grounded IMU simulation and graph encoding, achieving significant improvements in zero-shot activity recognition, cross-modal retrieval, and motion captioning across multiple datasets.