SMamDiff: Spatial Mamba for Stochastic Human Motion Prediction

TL;DR

SMamDiff introduces a novel single-stage diffusion model for human motion prediction that ensures spatial-temporal coherence, improving accuracy and efficiency over existing methods.

Contribution

The paper proposes SMamDiff, a spatial Mamba-based diffusion model with residual-DCT encoding and a joint-by-joint processing module for coherent, probabilistic human motion prediction.

Findings

Achieves state-of-the-art results on Human3.6M and HumanEva datasets.

Reduces latency and memory usage compared to multi-stage diffusion methods.

Ensures spatial-temporal coherence in single-stage probabilistic HMP.

Abstract

With intelligent room-side sensing and service robots widely deployed, human motion prediction (HMP) is essential for safe, proactive assistance. However, many existing HMP methods either produce a single, deterministic forecast that ignores uncertainty or rely on probabilistic models that sacrifice kinematic plausibility. Diffusion models improve the accuracy-diversity trade-off but often depend on multi-stage pipelines that are costly for edge deployment. This work focuses on how to ensure spatial-temporal coherence within a single-stage diffusion model for HMP. We introduce SMamDiff, a Spatial Mamba-based Diffusion model with two novel designs: (i) a residual-DCT motion encoding that subtracts the last observed pose before a temporal DCT, reducing the first DC component ($f=0$) dominance and highlighting informative higher-frequency cues so the model learns how joints move rather than where they are; and (ii) a stickman-drawing spatial-mamba module that processes joints in an ordered, joint-by-joint manner, making later joints condition on earlier ones to induce long-range, cross-joint dependencies. On Human3.6M and HumanEva, these coherence mechanisms deliver state-of-the-art results among single-stage probabilistic HMP methods while using less latency and memory than multi-stage diffusion baselines.