Diffusion Model-based Activity Completion for AI Motion Capture from Videos

TL;DR

This paper introduces a diffusion-model-based approach for AI motion capture that enables the generation of smooth, continuous human motion sequences beyond observed data, improving naturalness and coherence.

Contribution

It presents a novel diffusion model with a gate and position-time embedding modules for action completion in AI motion capture, addressing transition gaps in training data.

Findings

MDC-Net outperforms existing methods in ADE, FDE, MMADE

MDC-Net has a smaller model size than HumanMAC

MDC-Net produces more natural and coherent motions

Abstract

AI-based motion capture is an emerging technology that offers a cost-effective alternative to traditional motion capture systems. However, current AI motion capture methods rely entirely on observed video sequences, similar to conventional motion capture. This means that all human actions must be predefined, and movements outside the observed sequences are not possible. To address this limitation, we aim to apply AI motion capture to virtual humans, where flexible actions beyond the observed sequences are required. We assume that while many action fragments exist in the training data, the transitions between them may be missing. To bridge these gaps, we propose a diffusion-model-based action completion technique that generates complementary human motion sequences, ensuring smooth and continuous movements. By introducing a gate module and a position-time embedding module, our approach achieves competitive results on the Human3.6M dataset. Our experimental results show that (1) MDC-Net outperforms existing methods in ADE, FDE, and MMADE but is slightly less accurate in MMFDE, (2) MDC-Net has a smaller model size (16.84M) compared to HumanMAC (28.40M), and (3) MDC-Net generates more natural and coherent motion sequences. Additionally, we propose a method for extracting sensor data, including acceleration and angular velocity, from human motion sequences.