AI Paper Atlas

Research map / Video / Video generation

Unified Video Dense Prediction from Disjoint Data

Yihong Sun, Seoung Wug Oh, Jiahui Huang, Bharath Hariharan, Joon-Young Lee

ECCV 2026 · Video › Video generation · approach: Trajectory-guided motion control · 0 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
use diffusion-prior backbone instead of frozen DINOv3-H with DPT headsunified dense predictor · IIW albedo (OOD)WHDR 15%: 0.289 -> 0.207Improved
use diffusion-prior backbone instead of frozen DINOv3-H with DPT headsunified dense predictor · SAW shading (OOD)P@0.7: 87.8 -> 94.6Improved
use diffusion-prior backbone instead of frozen DINOv3-H with DPT headsunified dense predictor · Mapillary instance boundary (OOD)odsF: 55.0 -> 57.2Improved
use diffusion-prior backbone instead of frozen DINOv3-H with DPT headsunified dense predictor · NYUv2 depthAbsRel: 0.092 -> 0.059Improved
use specialist diffusion model instead of DINOv3-H under color jitteringspecialist · Hypersim val color jitter, depthAbsRel deviation: 0.034 -> 0.011Improved
use specialist diffusion model instead of DINOv3-H under color jitteringspecialist · Hypersim val color jitter, normalsmean error deviation: 5.8 -> 2.2Improved
distill all eight specialists into one unified backbone (cross-task regularization from depth)surface normal specialist · ScanNet surface normalsmean angular error: 15.4 -> 14.6Improved
add projector-only fine-tuning after latent l1 distillationUniD unified · Cityscapes semantic segmIoU: 37.8 -> 58.4Improved
add projector-only fine-tuning after latent l1 distillationUniD unified · ADE20K semantic seg (OOD)mIoU: 35.7 -> 44.8Improved
add video training (all videos as independent images otherwise), no inference memory (ablation)UniD unified · ScanNet video depth (temporal)AbsRel: 0.121 -> 0.109Improved
add video training, no inference memory (ablation)UniD unified · VIPSeg semantic temporal consistencymVC-4: 85.9 -> 86.5Improved
add extended self-attention inference memory [16,1] without video training (ablation)UniD unified · VIPSeg semantic temporal consistencymVC-4: 85.9 -> 89.1Improved

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