AI Paper Atlas

Research map / 3D vision / Depth and stereo

DAViD: Data-efficient and Accurate Vision Models from Synthetic Data

Fatemeh Saleh, Sadegh Aliakbarian, Charlie Hewitt, Lohit Petikam, Xian Xiao, Antonio Criminisi, Thomas J. Cashman, Tadas Baltrusaitis

ICCV 2025 · 3D vision › Depth and stereo · approach: Foundation models for monocular geometry

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 on SynthHuman instead of THuman2.0 renders (ablation)Large depth model · Goliath depthRMSE: 0.495 -> 0.253Improved
train on SynthHuman instead of RenderPeople renders (ablation)Large depth model · Goliath depthRMSE: 0.278 -> 0.253Improved
increase training set from 60K to 300K images (ablation)ViT-Base depth · Goliath depthRMSE: 0.324 -> 0.278Improved
increase training set from 60K to 300K images (ablation)ViT-Base depth · Hi4D depthRMSE: 0.101 -> 0.085Improved
scale backbone from ViT-Small to ViT-Large (ablation)depth model · Goliath depthRMSE: 0.310 -> 0.253Improved
scale backbone from ViT-Base to ViT-Large (ablation)depth model · Hi4D depthRMSE: 0.085 -> 0.072Improved
single multi-task Large model instead of three separate Large models (ablation)Large model · Goliath depthRMSE: 0.253 -> 0.270Hurt
single multi-task Large model instead of three separate Large models (ablation)Large model · PPM-100 mattingSAD: 78.17 -> 66.08Improved
multi-task Large (0.35B) instead of three Base models (3x0.12B) (ablation)Large model · Goliath depthRMSE: 0.278 -> 0.270Improved
blur the normal prediction output (ablation)Ours-Large normals · Goliath-FullBodymean angular error: 14.60 -> 14.52Improved
blur the normal prediction output (ablation)Sapiens-0.3B normals · Goliath-FullBodymean angular error: 15.72 -> 15.69No clear effect
fine-tune SynthHuman model on P3M-trainOurs matting · P3M-500-NPSAD: 14.83 -> 9.12Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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