atlas

Research map / Video / Video segmentation and tracking

A Neural Temporal Model for Human Motion Prediction

CVPR 2019 · Video › Video segmentation and tracking · approach: Physics-inspired motion prediction · 193 citations (OpenAlex)

Check your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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
Use the full two-level VTLN-RNN architecture, derivative features, and multi-objective loss instead of the ablated component variants (ablation)VTLN-RNN long-term motion synthesis · Human3.6M long-term motion synthesisperformanceHelped
Use the multi-objective loss rather than training without the proposed cost (ablation)VTLN-RNN long-term motion synthesis · Human3.6M long-term motion synthesisperformanceHelped
Add the two-level backward-planning VTLN-RNN architecture instead of a regular RNN (ablation)long-term motion synthesis model · Human3.6M long-term motion synthesisperformanceHelped
Append approximate joint-angle derivatives as input features (ablation)long-term motion synthesis model · Human3.6M long-term motion synthesisperformanceHelped
Use NPSS instead of MSE to rank long-term generated motion qualitymotion evaluation metrics · Long-term motion user study, all actionsuser agreement ratio: 0.4875 (39/80) -> 0.8125 (65/80)Helped
Use VGRU-d or GRU-d rather than MBR-long or VGRU-aclong-term motion synthesis · Human3.6M, all four actionsNPSSHelped
Use short-term models for long-horizon synthesis rather than long-term modelsmotion synthesis models · Human3.6M, 2–4 secondsNPSSHurt

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