Progressively Generating Better Initial Guesses Towards Next Stages for High-Quality Human Motion Prediction

TLDR

A good initial guess of future poses is crucial for improving human motion forecasting accuracy. The study proposes a high‑quality human motion prediction method that generates progressively better initial guesses across multiple stages to improve forecasting. The method uses a multi‑stage framework with an initial‑prediction network followed by a formal‑prediction network, employing Spatial and Temporal Dense Graph Convolutional Networks to extract global spatiotemporal features. The approach outperforms prior methods by 6–7 % on Human3.6M, 5–10 % on CMU‑MoCap, and 13–16 % on 3DPW. Code is available at https://github.com/705062791/PGBIG.

Abstract

This paper presents a high-quality human motion pre-diction method that accurately predicts future human poses given observed ones. Our method is based on the observation that a good "initial guess" of the future poses is very helpful in improving the forecasting accuracy. This mo-tivates us to propose a novel two-stage prediction frame-work, including an init-prediction network that just computes the good guess and then a formal-prediction network that predicts the target future poses based on the guess. More importantly, we extend this idea further and design a multi-stage prediction framework where each stage pre-dicts initial guess for the next stage, which brings more performance gain. To fulfill the prediction task at each stage, we propose a network comprising Spatial Dense Graph Convolutional Networks (S-DGCN) and Temporal Dense Graph Convolutional Networks (T-DGCN). Alternatively executing the two networks helps extract spatiotem-poral features over the global receptive field of the whole pose sequence. All the above design choices cooperating together make our method outperform previous approaches by large margins: 6%-7% on Human3.6M, 5%-10% on CMU-MoCap, and 13%-16% on 3DPW. Code is available at https://github.com/705062791/PGBIG.