Universal Skeleton Understanding via Differentiable Rendering and MLLMs

TL;DR

SkeletonLLM enables universal understanding of human skeleton data by translating skeleton sequences into visual tokens for multimodal large language models, enhancing reasoning and generalization across diverse formats.

Contribution

The paper introduces SkeletonLLM with DrAction, a differentiable renderer, and a cooperative training strategy for effective skeleton understanding in MLLMs, addressing format heterogeneity.

Findings

Strong generalization in open-vocabulary action recognition

Effective motion captioning and question answering across skeleton formats

End-to-end differentiable pipeline improves task-specific visual token generation

Abstract

Multimodal large language models (MLLMs) exhibit strong visual-language reasoning, yet cannot process structured, non-visual data such as human skeletons. Existing methods either compress skeleton dynamics into lossy feature vectors for text alignment, or quantize motion into discrete tokens that generalize poorly across heterogeneous skeleton formats. We present SkeletonLLM, which achieves universal skeleton understanding by translating arbitrary skeleton sequences into the MLLM's native visual modality. At its core is DrAction, a differentiable, format-agnostic renderer that converts skeletal kinematics into compact image sequences. Because the pipeline is end-to-end differentiable, MLLM gradients can directly guide the rendering to produce task-informative visual tokens. To further enhance reasoning capabilities, we introduce a cooperative training strategy: Causal Reasoning Distillation transfers structured, step-by-step reasoning from a teacher model, while Discriminative Finetuning sharpens decision boundaries between confusable actions. SkeletonLLM demonstrates strong generalization \revise{in open-vocabulary action recognition, while its learned reasoning capabilities naturally extend to motion captioning and question answering across heterogeneous skeleton formats} -- suggesting a viable path for applying MLLMs to non-native modalities. Code: https://github.com/wangzy01/SkeletonLLM.