Paper page - HumanSkill-4D: Training-Free Language Querying of Animatable Gaussian Avatars in Space and Time

Source: https://huggingface.co/papers/2610.24452

Abstract

HumanSkill-4D enables language-driven selection of avatar regions and motion intervals via shared multimodal reasoning and geometric operators without training.

We introduceavatar referring segmentation, which selects regions of animatableGaussian avatarsfrom language queries. A query can also specify a pose or motion to select matching intervals. This lets users interact with avatars for local editing and motion inspection. However, making these selections with existing language Gaussian splatting methods poses two challenges: adjacent parts and garments are hard to separate and may be referred to indirectly (region ambiguity), while fine-grained motions can look alike in RGB (motion ambiguity). We address these challenges with HumanSkill-4D, a training-free framework that shares amultimodal large language model(MLLM) andgeometric operatorsacross avatars. For each avatar, asemantic memorylinks body-part and garment categories to persistent Gaussian regions and appearance descriptions. The MLLM uses category names and descriptions to select categories by appearance, relation, or function. To find matching intervals, it composes operators into programs that check query conditions on driving joint trajectories stored in amotion memory. These category choices and operator outputs make selections traceable, while composing or adding operators supports new motion conditions. To evaluate this task, we introduce three human-annotated benchmarks and show that HumanSkill-4D outperforms the baselines in part segmentation and 3D and4D referring segmentation, with compact memories and only minutes of setup per avatar.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2610\.24452

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/2610.24452 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.24452 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.24452 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.