AI Paper Atlas

Research map / 3D vision / 3D generation

Self-Supervised Pre-Training With Masked Shape Prediction for 3D Scene Understanding

Li Jiang, Zetong Yang, Shaoshuai Shi, Vladislav Golyanik, Dengxin Dai, Bernt Schiele

CVPR 2023 · 3D vision › 3D generation · approach: 2D-to-3D pretrained representations · 13 citations (OpenAlex)

Read the paper

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
Apply MSP pre-trainingEQ-Net · ScanNet semantic segmentationmIoU (%): 73.64 -> 75.57Improved
Apply MSP pre-trainingEQ-Net with GroupFree L6,O256 · ScanNet v2 indoor detectionAP50: 51.1 -> 53.7Improved
Apply MSP pre-trainingEQ-Net with CenterPoint · Waymo Level 2, 20% training datamAPH Avg.: 63.51 -> 64.26Improved
Add shape context and deep shape feature together rather than using either alone (ablation)MSP-SA · ScanNet semantic segmentationmIoU (%): 75.05 / 74.92 -> 75.42Improved
Add auxiliary color prediction to the context-enhanced target (ablation)MSP-SA · ScanNet semantic segmentationmIoU (%): 75.42 -> 75.57Improved
Use MSP-SA with 10k sampled keypoints instead of MSP-CA or MSP-CA++ (ablation)MSP pre-training · ScanNet semantic segmentationmIoU (%): 74.85 / 75.05 -> 75.57Improved
Increase MSP-SA keypoint sampling from 10k to 40k (ablation)MSP pre-training · ScanNet semantic segmentationmIoU (%): 75.57 -> 74.49Hurt
Set masking ratio to 80% instead of 60% (ablation)MSP-SA · ScanNet semantic segmentationmIoU (%): 75.57 -> 74.34Hurt
Set block size to 0.5m instead of 0.3m (ablation)MSP-SA · ScanNet semantic segmentationmIoU (%): 75.57 -> 74.62Hurt
Replace EQ-Net with SparseConvNet as feature extractor (ablation)MSP pre-training · ScanNet semantic segmentationmIoU (%): 75.57 -> 74.13Hurt
Use local point set rather than context-enhanced shape target (ablation)MSP-SA · ScanNet semantic segmentationmIoU (%): 75.57 -> 73.71Hurt
Use color alone rather than context-enhanced shape target (ablation)MSP-SA · ScanNet semantic segmentationmIoU (%): 75.57 -> 73.98Hurt

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