FutureHuman3D: Forecasting Complex Long-Term 3D Human Behavior from Video Observations

TL;DR

FutureHuman3D introduces a weakly supervised generative model that forecasts complex long-term 3D human behaviors from 2D video data, enabling realistic 3D motion prediction without extensive 3D annotations.

Contribution

The paper presents a novel method combining 2D supervision with 3D generation, jointly predicting actions and poses to improve long-term human behavior forecasting.

Findings

Joint action and pose prediction outperform individual models.

Method effectively forecasts complex multi-action sequences.

Approach requires only 2D data at inference, reducing data collection costs.

Abstract

We present a generative approach to forecast long-term future human behavior in 3D, requiring only weak supervision from readily available 2D human action data. This is a fundamental task enabling many downstream applications. The required ground-truth data is hard to capture in 3D (mocap suits, expensive setups) but easy to acquire in 2D (simple RGB cameras). Thus, we design our method to only require 2D RGB data at inference time while being able to generate 3D human motion sequences. We use a differentiable 2D projection scheme in an autoregressive manner for weak supervision, and an adversarial loss for 3D regularization. Our method predicts long and complex human behavior sequences (e.g., cooking, assembly) consisting of multiple sub-actions. We tackle this in a semantically hierarchical manner, jointly predicting high-level coarse action labels together with their low-level fine-grained realizations as characteristic 3D human poses. We observe that these two action representations are coupled in nature, and joint prediction benefits both action and pose forecasting. Our experiments demonstrate the complementary nature of joint action and 3D pose prediction: our joint approach outperforms each task treated individually, enables robust longer-term sequence prediction, and improves over alternative approaches to forecast actions and characteristic 3D poses.