AI Paper Atlas

Research map / 3D vision / 3D generation

Towards Scalable Spatial Intelligence via 2D-to-3D Data Lifting

Xingyu Miao, Haoran Duan, Quanhao Qian, Jiuniu Wang, Yang Long, Ling Shao, Deli Zhao, Ran Xu, Gongjie Zhang

ICCV 2025 · 3D vision › 3D generation · approach: Open-vocabulary 3D scene understanding · 6 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
pretrain on COCO-3D before fine-tuning on ScanNetUni3D+Mask3D · ScanNet instance segmentationmAP (%): 24.30 -> 28.64Improved
pretrain on COCO-3D before fine-tuning on ScanNetUni3D+Mask3D · ScanNet instance segmentationmAP@0.25: 67.28 -> 67.33No clear effect
pretrain on COCO-3D before fine-tuning on ScanNetSpUNet · ScanNet semantic segmentationmIoU (%): 31.09 -> 62.48Improved
pretrain on COCO-3D before fine-tuning on ScanNetPTv2 · ScanNet semantic segmentationmIoU (%): 51.04 -> 55.81Improved
pretrain on COCO-3D before fine-tuning on ScanNetUni3D · ScanNet semantic segmentationmIoU (%): 52.14 -> 55.83Improved
pretrain on COCO-3D before fine-tuning on ScanNetTGNN · Referring instance segmentation (ScanNet)mIoU (%): 26.10 -> 32.47Improved
pretrain on COCO-3D before fine-tuning on ScanNetTGNN · Referring instance segmentation (ScanNet)Acc@0.5: 29.00 -> 37.12Improved
pretrain on COCO-3DLL3DA · ScanQA (val)CIDEr: 76.8 -> 85.2Improved
pretrain on COCO-3DLL3DA · ScanRefer dense captioningC@0.5: 62.98 -> 67.04Improved
pretrain on COCO-3DLL3DA · Nr3D dense captioningC@0.5: 23.94 -> 25.20Improved
train on COCO-3D only and test zero-shot on ScanNetTGNN · Referring instance segmentationmIoU (%): 26.10 -> 10.10Hurt

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