AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

A Closer Look at Invariances in Self-Supervised Pre-training for 3D Vision

Lanxiao Li, Michael Heizmann

ECCV 2022 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Invariance-based self-supervised learning · 22 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
pre-train with DPCo (depth-point contrast) instead of training from scratchVoteNet (PointNet++) · ScanNet detectionAP50: 37.6 -> 41.5Improved
contrast a 2D format (depth map) with point clouds instead of point-vs-voxel (3D-3D) (ablation)VoteNet (PointNet++) · ScanNet detectionAP25: 62.8 -> 64.2Improved
add voxel-point format contrast (PVCo) over augmentation-only PPCo (ablation)VoteNet (PointNet++) · SUN RGB-D detectionAP25: 58.6 -> 59.3No clear effect
use explicit perspective-invariance (PointContrast) instead of augmentation-only PPCo (ablation)VoteNet (PointNet++) · SUN RGB-D and ScanNet detectionAP50: 34.9 -> 34.1 (SUN RGB-D), 39.5 -> 38.1 (ScanNet)Mixed
replace RGB image modality (IPCo) with depth map format (DPCo) (ablation)VoteNet (PointNet++) · SUN RGB-D detectionAP50: 35.5 -> 35.6No clear effect
add local correspondence only (DPCo) vs from scratch (ablation)VoteNet (PointNet++) · ScanNet detectionAP25: 60.0 -> 64.2Improved
add global correspondence only (DPCo) vs from scratch (ablation)VoteNet (PointNet++) · ScanNet detectionAP25: 60.0 -> 63.8Improved
combine local and global correspondence (DPCo) instead of local only (ablation)VoteNet (PointNet++) · SUN RGB-D detectionAP50: 35.1 -> 35.6Improved
DPCo pre-training with only 5% of ScanNet training dataVoteNet (PointNet++) · ScanNet detection (5% data)AP25: 13.3 -> 36.5Improved
DPCo pre-training with only 5% of ScanNet training dataVoteNet (PointNet++) · ScanNet detection (5% data)AP50: 2.4 -> 14.4Improved
pre-train depth map encoder with depth-depth contrast (DDCo) vs from scratch2.5D-VoteNet · SUN RGB-D detectionAP25: 60.8 -> 56.0Hurt
add point cloud contrast (DPCo) to depth map encoder pre-training instead of depth-depth only (ablation)2.5D-VoteNet · SUN RGB-D detectionAP25: 56.0 -> 61.4Improved

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