KHMP: Frequency-Domain Kalman Refinement for High-Fidelity Human Motion Prediction

2026-03-22

AI Summary

This work proposes a frequency-domain adaptive Kalman refinement mechanism to address temporal jitter and discontinuities in stochastic human motion prediction, which are often caused by high-frequency noise. The approach first maps motion sequences into the frequency domain via the Discrete Cosine Transform (DCT), modeling high-frequency coefficients as noisy signals. An adaptive Kalman filter is then introduced to dynamically adjust noise parameters based on signal-to-noise ratio, effectively suppressing noise while preserving fine motion details. Furthermore, biomechanical constraints—including temporal smoothness and joint-angle limits—are integrated to ensure physical plausibility of the generated motions. Experiments on the Human3.6M and HumanEva-I datasets demonstrate that the method significantly reduces jitter, produces smooth and physically consistent motion sequences, and achieves state-of-the-art prediction accuracy.

Technology Category

• Computer Vision: Motion & Tracking
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Stochastic human motion prediction aims to generate diverse, plausible futures from observed sequences. Despite advances in generative modeling, existing methods often produce predictions corrupted by high-frequency jitter and temporal discontinuities. To address these challenges, we introduce KHMP, a novel framework featuring an adaptiveKalman filter applied in the DCT domain to generate high-fidelity human motion predictions. By treating high-frequency DCT coefficients as a frequency-indexed noisy signal, the Kalman filter recursively suppresses noise while preserving motion details. Notably, its noise parameters are dynamically adjusted based on estimated Signal-to-Noise Ratio (SNR), enabling aggressive denoising for jittery predictions and conservative filtering for clean motions. This refinement is complemented by training-time physical constraints (temporal smoothness and joint angle limits) that encode biomechanical principles into the generative model. Together, these innovations establish a new paradigm integrating adaptive signal processing with physics-informed learning. Experiments on the Human3.6M and HumanEva-I datasets demonstrate that KHMP achieves state-of-the-art accuracy, effectively mitigating jitter artifacts to produce smooth and physically plausible motions.

Problem

• human motion prediction
• high-frequency jitter
• temporal discontinuities
• motion fidelity
• stochastic prediction

Innovation

• Kalman filter
• DCT domain
• adaptive denoising
• physics-informed learning
• human motion prediction