AI Paper Atlas

Research map / 3D vision / Depth and stereo

Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets From 3D Scans

Ainaz Eftekhar, Alexander Sax, Jitendra Malik, Amir Zamir

ICCV 2021 · 3D vision › Depth and stereo · approach: Self-supervised monocular depth · 291 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
train depth model on Omnidata starter data instead of MiDaS dataDPT-Hybrid · NYU zero-shot depthL1 error: 0.3838 -> 0.2878Improved
train depth model on Omnidata starter data instead of MiDaS dataDPT-Hybrid · OASIS zero-shot depthL1 error: 0.8057 -> 0.7901Improved
train normals UNet on Omnidata instead of OASIS-trained HourglassUNet · OASIS surface normalsAUC_p: 0.5786 -> 0.6758Improved
train normals UNet on Omnidata instead of OASIS-trained HourglassUNet · OASIS surface normalsmean angular error: 23.91 -> 24.87Hurt
train UNet on Omnidata instead of SunCGUNet · OASIS surface normalsmean angular error: 35.42 -> 24.87Improved
add normals as extra input cue (ablation)HRNet-18 · Semantic segmentation, Replica to Hypersim (GT cues)cross-entropy: 5.87 -> 4.47Improved
add 2D edges, z-depth and 3D keypoints cues to the input (ablation)HRNet-18 · Semantic segmentation, Hypersim (GT cues)cross-entropy: 4.47 -> 3.86Improved
add 3D edges cue to the input (ablation)HRNet-18 · Semantic segmentation, Taskonomy (GT cues)cross-entropy: 6.12 -> 6.75Hurt
train on all starter data instead of a single component datasetSurface normal UNet · Inter-dataset transferharmonic mean L1: 8.24 -> 6.11Improved
use multi-task learning methods (MTAN, cross-stitch) over single taskMTL methods · Omnidata semantic segmentationIoU: 85.12 -> 83.00Hurt
apply MTAN to MTL baselineMTL baseline · Omnidata semantic segmentationIoU: 81.82 -> 83.00Improved

Later papers that compared with this method

3 later accepted papers report Omnidata v2 as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.

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