AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Embed Me if You Can: A Geometric Perceptron

Pavlo Melnyk, Michael Felsberg, Mårten Wadenbäck

ICCV 2021 · Point clouds › Shape and surface completion · approach: Unsupervised shape correspondence · 6 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 MLGP point-wise embedding instead of MLP hidden layerMLGP vs MLP · Tetris main dataset, noise a=0.0test accuracy (%): 71.7 -> 91.8Improved
use MLGP point-wise embedding instead of MLHP embeddingMLGP vs MLHP · Tetris main dataset, noise a=0.0test accuracy (%): 92.5 -> 91.8No clear effect
use MLGP instead of MLHP under noise a=0.2MLGP vs MLHP · Tetris main dataset, noise a=0.2test accuracy (%): 68.2 -> 69.7Improved
use MLGP instead of MLHP under noise a=0.2, top-10 by validationMLGP vs MLHP · Tetris main dataset, noise a=0.2test accuracy (%): 72.2 -> 79.6Improved
use MLGP instead of MLHP on theta-split without noiseMLGP vs MLHP · Tetris theta-split, noise a=0.0test accuracy (%): 87.6 -> 87.9Improved
use MLGP instead of MLHP on theta-split noise a=0.1, top-10MLGP vs MLHP · Tetris theta-split, noise a=0.1test accuracy (%): 83.8 -> 87.5Improved
train/test on disjoint rotation ranges (theta-split) for MLPMLP · Tetris theta-split vs main, noise a=0.0test accuracy (%): 71.7 -> 55.3Hurt
train/test on disjoint rotation ranges (theta-split) for MLGPMLGP · Tetris theta-split vs main, noise a=0.0test accuracy (%): 91.8 -> 87.9Hurt
train/test on disjoint rotation ranges (theta-split) for MLHPMLHP · Tetris theta-split vs main, noise a=0.0test accuracy (%): 92.5 -> 87.6Hurt
remove hidden-layer activation functions (identity) for MLGP and MLHP (ablation)MLGP and MLHP · Tetristest accuracy (%)Improved
apply the same rigid transformation to geometric neuron weights as to the test data (ablation)MLGP · Tetris isometry testtest accuracy (%): 86.09 -> 91.98Improved
transform only the data but not the weights (ablation)MLGP · Tetris isometry testtest accuracy (%): 91.98 -> 86.09Hurt

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