AI Paper Atlas

Research map / 3D vision / 3D generation

Feed-Forward SceneDINO for Unsupervised Semantic Scene Completion

Aleksandar Jevtić, Christoph Reich, Felix Wimbauer, Oliver Hahn, Christian Rupprecht, Stefan Roth, Daniel Cremers

ICCV 2025 · 3D vision › 3D generation · approach: Semantic scene completion · 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 the downsampler for high-resolution target features (vs bilinear upsampling + augmentation) (ablation)SceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 6.82 -> 8.00Improved
add the feature smoothness loss (ablation)SceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 6.83 -> 8.00Improved
add positional-encoding decomposition (ablation)SceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 7.26 -> 8.00Improved
use estimated ORB-SLAM3 poses instead of KITTI-360 poses (ablation)SceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 8.00 -> 7.88No clear effect
switch target features from DINO to DINOv2 (ablation)SceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 8.00 -> 9.08Improved
add 3D feature distillation (vs clustering raw features) (ablation)SceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 6.39 -> 8.00Improved
add kNN-correlation loss (ablation)SceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 6.65 -> 8.00Improved
add neighborhood sampling around center points (ablation)SceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 7.03 -> 8.00Improved
use 3D sampling instead of 5-crop sampling (ablation)SceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 7.53 -> 8.00Improved
lift DINO features into a 3D feature field with distillation vs S4C + STEGO baselineSceneDINO · SSCBench-KITTI-360 (51.2m)mIoU: 6.60 -> 8.00Improved
add 2D-supervised linear probing instead of unsupervised probingSceneDINO (DINO) · SSCBench-KITTI-360 (51.2m)mIoU: 8.00 -> 9.34Improved
use FiT3D 3D-refined features instead of DINO in STEGOSTEGO · KITTI-360 2D segmentationmIoU: 23.57 -> 22.29Hurt

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