AI Paper Atlas

Research map / Detection and segmentation / Weakly supervised detection and segmentation

TWIST: Two-Way Inter-Label Self-Training for Semi-Supervised 3D Instance Segmentation

Ruihang Chu, Xiaoqing Ye, Zhengzhe Liu, Xiao Tan, Xiaojuan Qi, Chi-Wing Fu, Jiaya Jia

CVPR 2022 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Semi-supervised instance segmentation · 27 citations (OpenAlex)

Check 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 TWIST self-training with semantic and instance pseudo labels3D instance segmentation · ScanNet v2 validation, 1% labelsmAP: 5.1 -> 9.6Improved
Initialize with CSC-pretrained weightsTWIST · ScanNet v2 validation, 1% labelsmAP: 9.6 -> 11.5Improved
Add naive-threshold pseudo semantic labels (ablation)Supervised-only 3D instance segmentation · ScanNet v2 validation, 5% labelsmAP: 18.2 -> 20.9Improved
Add naive-threshold pseudo offset vectors (ablation)Supervised-only 3D instance segmentation · ScanNet v2 validation, 5% labelsmAP: 18.2 -> 19.7Improved
Use both naive-threshold semantic and offset pseudo labels (ablation)Supervised-only 3D instance segmentation · ScanNet v2 validation, 5% labelsmAP: 20.9 -> 21.3Improved
Generate pseudo labels at proposal level (ablation)TWIST self-training · ScanNet v2 validation, 5% labelsmAP: 21.3 -> 24.8Improved
Add proposal re-correction for object-level denoising (ablation)Proposal-based pseudo-label update · ScanNet v2 validation, 5% labelsmAP: 24.8 -> 27.0Improved
Remove offset-to-semantic interaction and use point-wise thresholding (ablation)TWIST · ScanNet v2 validation, 5% labelsmAP: 27.0 -> 22.5Hurt
Remove semantic-to-offset interaction using class-agnostic clustering (ablation)TWIST · ScanNet v2 validation, 5% labelsmAP: 27.0 -> 21.7Hurt
Use backbone features instead of RGB features for re-correction (ablation)Proposal re-correction module · ScanNet v2 validation, 5% labelsmAP: 22.3 -> 25.6Improved
Add geometric transformation augmentation (ablation)Proposal re-correction module · ScanNet v2 validation, 5% labelsmAP: 25.6 -> 25.9Improved
Add geometric mutation augmentation (ablation)Proposal re-correction module · ScanNet v2 validation, 5% labelsmAP: 25.6 -> 26.6Improved

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