AI Paper Atlas

Research map / 3D vision / Human pose estimation

Contextually Plausible and Diverse 3D Human Motion Prediction

Sadegh Aliakbarian, Fatemeh Saleh, Lars Petersson, Stephen Gould, Mathieu Salzmann

ICCV 2021 · 3D vision › Human pose estimation · approach: Human motion prediction · 49 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
condition decoder with extended reparameterization on z_c instead of concatenating z_c (ablation)CVAE decoder · Human3.6Mdiversity: 1.28 -> 3.35Improved
condition decoder with extended reparameterization on z_c instead of concatenating z_c (ablation)CVAE decoder · Human3.6Mcontext: 0.40 -> 0.51Improved
replace deterministic h_t decoder concatenation with stochastic z_c reparameterization (ablation)CVAE decoder · Human3.6Mdiversity: 0.24 -> 3.12Improved
replace deterministic h_t decoder concatenation with stochastic z_c reparameterization (ablation)CVAE decoder · Human3.6Mcontext: 0.43 -> 0.54Improved
use deterministic h_t instead of stochastic z_c to condition the encoder (ablation)LCP-VAE encoder · Human3.6Mcontext: 0.51 -> 0.54Improved
use deterministic h_t instead of stochastic z_c to condition the encoder (ablation)LCP-VAE encoder · Human3.6Mdiversity: 3.35 -> 3.12Hurt
learn a conditional prior (LCP-VAE) instead of DLowstochastic motion prediction · Human3.6Mcontext: 0.29 -> 0.54Improved
learn a conditional prior (LCP-VAE) instead of DLowstochastic motion prediction · Human3.6Mdiversity: 4.71 -> 3.12Hurt
learn a conditional prior (LCP-VAE) instead of Mix-and-Matchstochastic motion prediction · Human3.6Mcontext: 0.37 -> 0.54Improved
learn a conditional prior (LCP-VAE) instead of DLowstochastic motion prediction · CMU MoCapdiversity: 2.90 -> 2.36Hurt
learn a conditional prior (LCP-VAE) instead of DLowstochastic motion prediction · Human3.6M (best of K=50 sampled), Walking 80 msmean angle error: 0.31 -> 0.22Improved
LCP-VAE versus Mix-and-Match on long horizonstochastic motion prediction · Human3.6M (best of K=50 sampled), Walking 1000 msmean angle error: 0.68 vs 0.69No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy