AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

General 3D Vision-Language Model With Fast Rendering and Pre-Training Vision-Language Alignment

Kangcheng Liu, Yong‐Jin Liu, Baoquan Chen

TPAMI 2025 · Detection and segmentation › Open-set and few-shot detection · approach: Open-vocabulary 3D and world perception · 5 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
add hierarchical vision-language pre-training (Pr) (ablation)WS3D++ · ScanNet semantic segmentation (1% labels)mIoU: 52.6 -> 58.8Improved
add hierarchical vision-language pre-training (Pr) (ablation)WS3D++ · ScanNet semantic segmentation (100% labels)mIoU: 81.7 -> 88.8Improved
add pre-training (Pr) (ablation)WS3D++ · S3DIS semantic segmentation (1% labels)mIoU: 48.6 -> 55.8Improved
add pre-training (Pr) (ablation)WS3D++ · SemanticKITTI semantic segmentation (1% labels)mIoU: 46.8 -> 51.6Improved
add pre-training (Pr) (ablation)WS3D++ · S3DIS semantic segmentation (100% labels)mIoU: 82.0 -> 82.9Improved
add fine-tuning modules (region-level energy and contrastive losses) to the pre-trained model (ablation)WS3D++ · ScanNet semantic segmentation (1% labels)mIoU: 43.9 -> 58.8Improved
add pre-training (Pr) (ablation)WS3D++ · ScanNet instance segmentation (1% labels)AP@50: 45.9 -> 48.2Improved
add pre-training (Pr) (ablation)WS3D++ · ScanNet instance segmentation (100% labels)AP@50: 58.8 -> 73.9Improved
add pre-training (Pr) (ablation)WS3D++ · SUN RGB-D 3D detection (5% labels, inductive)mAP: 52.7 -> 56.9Improved
add pre-training (Pr) (ablation)WS3D++ · ScanNet-V2 3D detection (5% labels, inductive)mAP: 55.3 -> 66.6Improved
use transductive training (test set used for training) instead of inductive (ablation)WS3D++ · ScanNetv2 semantic segmentation (20% labels)mIoU: 71.8 -> 85.8Improved
use WS3D++ instead of GPCsemantic segmentation · ScanNet semantic segmentation (1% labels)mIoU: 46.6 -> 52.6Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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