AI Paper Atlas

Research map / 3D vision / Human pose estimation

MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation

Xucong Zhang, Yusuke Sugano, Mario Fritz, Andreas Bulling

TPAMI 2017 · 3D vision › Human pose estimation · approach: Learned camera pose estimation · 631 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
use GazeNet instead of the prior state-of-the-art appearance methodGazeNet · MPIIGaze (cross-dataset, train UT Multiview)mean angular error (degrees): 13.9 -> 10.8Improved
use GazeNet instead of the prior state-of-the-art methodGazeNet · EYEDIAP (cross-dataset, train UT Multiview)mean angular error (degrees): 8% on EYEDIAP (0.9 degrees)Improved
use 60x36 input instead of 15x9 pixels (ablation)GazeNet · MPIIGaze (cross-dataset, train UT Multiview)mean angular error (degrees): 11.4 -> 10.8Improved
use manually annotated facial landmarks instead of detected onesGazeNet · MPIIGaze+ (cross-dataset, train UT Multiview)mean angular error (degrees): 10.8 -> 9.8Improved
pretrain on synthetic data then fine-tune on UT MultiviewGazeNet+ · MPIIGaze+ (cross-dataset)mean angular error (degrees): 9.8 -> 9.1Improved
restrict training gaze/head-pose distribution to match target (UT subset)GazeNet+ · MPIIGaze+ (cross-dataset)mean angular error (degrees): 9.8 -> 8.0Improved
use synthetic training data with matched gaze rangeGazeNet+ · MPIIGaze+ (cross-dataset)mean angular error (degrees): 8.0 -> 7.3Improved
add detected pupil centre feature (ablation)GazeNet+ · MPIIGaze+ (cross-person, leave-one-person-out)mean angular error (degrees): 5.4 -> 5.2No clear effect
add manually annotated pupil centre feature (ablation)GazeNet+ · MPIIGaze+ (cross-person, leave-one-person-out)mean angular error (degrees): 5.4 -> 5.0No clear effect
fuse both eyes by averaging 3D gaze vectors instead of averaging single-eye estimatesGazeNet+ · MPIIGaze+ (both eyes)mean angular error (degrees): 9.8 -> 7.2Improved
use generic instead of personal 3D face model for head pose (ablation)GazeNet+ · MPIIGaze+ (cross-dataset, train UT Multiview)mean angular error (degrees): 9.8 -> 9.7No clear effect
train and test within the same dataset instead of cross-datasetGazeNet+ · MPIIGaze+mean angular error (degrees): 9.1 -> 5.4Improved

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