English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present a generative approach to forecast long-term future human behaviorin 3D, requiring only weak supervision from readily available 2D human actiondata. This is a fundamental task enabling many downstream applications. Therequired ground-truth data is hard to capture in 3D (mocap suits, expensivesetups) but easy to acquire in 2D (simple RGB cameras). Thus, we design ourmethod to only require 2D RGB data while being able to generate 3D human motionsequences. We use a differentiable 2D projection scheme in an autoregressivemanner for weak supervision, and an adversarial loss for 3D regularization. Ourmethod predicts long and complex behavior sequences (e.g. cooking, assembly)consisting of multiple sub-actions. We tackle this in a semanticallyhierarchical manner, jointly predicting high-level coarse action labelstogether with their low-level fine-grained realizations as characteristic 3Dhuman poses. We observe that these two action representations are coupled innature, and joint prediction benefits both action and pose forecasting. Ourexperiments demonstrate the complementary nature of joint action and 3D poseprediction: our joint approach outperforms each task treated individually,enables robust longer-term sequence prediction, and outperforms alternativeapproaches to forecast actions and characteristic 3D poses.

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