AI Paper Atlas

Research map / 3D vision / Human pose estimation

Modeling Human Motion with Quaternion-Based Neural Networks

Dario Pavllo, Christoph Feichtenhofer, Michael Auli, David Grangier

IJCV 2019 · 3D vision › Human pose estimation · approach: Human motion prediction · 225 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
model velocities instead of absolute rotations (ablation)QuaterNet (RNN) · Human3.6M short-term, proposed protocol, average, 80 msmean angle error: 0.45 -> 0.37Improved
model velocities instead of absolute rotations (ablation)QuaterNet (RNN) · Human3.6M short-term, proposed protocol, average, 400 msmean angle error: 1.17 -> 1.14Improved
add scheduled sampling (vs teacher forcing) (ablation)QuaterNet vel. (RNN) · Human3.6M short-term, proposed protocol, average, 400 msmean angle error: 1.23 -> 1.14Improved
add scheduled sampling (vs teacher forcing) (ablation)QuaterNet vel. (RNN) · Human3.6M short-term, proposed protocol, average, 80 msmean angle error: 0.35 -> 0.37Hurt
add scheduled sampling (vs teacher forcing) (ablation)QuaterNet CNN vel. · Human3.6M short-term, proposed protocol, average, 400 msmean angle error: 1.30 -> 1.19No clear effect
use recurrent instead of convolutional architectureQuaterNet vel. · Human3.6M short-term, proposed protocol, average, 400 msmean angle error: 1.19 -> 1.14Improved
use recurrent instead of convolutional architectureQuaterNet abs. · Human3.6M short-term, proposed protocol, average, 400 msmean angle error: 1.55 -> 1.17Improved
evaluate with 128 samples per sequence instead of 4 (ablation)zero-velocity baseline · Human3.6M average error variance (25th-75th quantile)mean angle error: 0.395 -- 0.435 -> 0.413 -- 0.420Improved
model absolute rotations instead of velocities for long horizon (ablation)QuaterNet · Human3.6M long-term (up to 1 s)mean angle errorImproved
use forward-kinematics position loss instead of angle loss (ablation)QuaterNet · Locomotion generationposition errorImproved
use quaternions instead of Euler angles or axis-angle (ablation)QuaterNet · Locomotion generation (long-term)position and velocity errorImproved
increase conditioning length beyond 10-20 frames (ablation)QuaterNet · Human3.6M Walking, 80 msangle errorNo 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