AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Temporally-Coherent Surface Reconstruction via Metric-Consistent Atlases

Jan Bednarik, Vladimir G. Kim, Siddhartha Chaudhuri, Shaifali Parashar, Mathieu Salzmann, Pascal Fua, Noam Aigerman

ICCV 2021 · Point clouds › Shape and surface completion · approach: Unsupervised shape correspondence · 8 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
sample training pairs from a time window of neighbors instead of random pairs (ablation)OUR · AMA cranem_sL2: 96.82 -> 66.63Improved
sample training pairs from a time window of neighbors instead of random pairs (ablation)OUR · AMA cranemPCK AUC: 77.08 -> 80.24Improved
sample training pairs from a time window of neighbors instead of random pairs (ablation)OUR · AMA craneChamfer distance: 0.30 -> 0.31Hurt
add metric consistency loss (weight 0.1 vs 1e-3) (ablation)OUR · AMA cranem_sL2: 232.5 -> 66.6Improved
add metric consistency loss (weight 1e-2 vs 1e-4) (ablation)OUR · ANIM catm_sL2: 13.2 -> 9.8Improved
raise metric weight from 0.1 to 1e3 (ablation)OUR · DFAUST jacksm_sL2: 28.0 -> 194.2Hurt
replace AtlasNet with metric-consistent atlas (add metric loss)AtlasNet · ANIMm_sL2: 18.40 -> 11.93Improved
replace AtlasNet with metric-consistent atlas (add metric loss)AtlasNet · AMAm_sL2: 86.80 -> 57.12Improved
replace AtlasNet with metric-consistent atlas (add metric loss)AtlasNet · AMAChamfer distance: 0.30 -> 0.32Hurt
replace AtlasNet with metric-consistent atlas (add metric loss)AtlasNet · DFAUSTm_sL2: 31.74 -> 19.81Improved
train CC with random triplets instead of knn (ablation)CC · DFAUST jumping_jacksm_sL2: 105.57 -> 32.74Improved
use random nrICP matching instead of propagate_simple (ablation)nrICP · AMA cranem_sL2: 211.11 -> 172.55Improved

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