Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data | VitaVision
PaperCVPR 2024 · 2024

Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data

Lihe YangBingyi KangZilong HuangXiaogang XuJiashi FengHengshuang Zhao
arXiv 2401.10891DOIView in graph
1page built on it
2more pages cite it
1narrative
0registry papers cite it

In the Atlas

ModelPrimary source of
Depth Anything

A monocular relative-depth foundation model that scales training to 62M unlabeled images via a teacher pseudo-labeling data engine, a CutMix challenge for the student, and a DINOv2 semantic-feature alignment loss.

Cited as a reference by 2 pages, grouped by domain

Depth 2
Depth Anything V2Monocular Depth Estimation

Appears in 1 narrative

NarrativeDepth Becomes General Geometry

Lineage

Papers in the Atlas registry only · one dot per paper
Built upon by · 0
’24
Builds on · 0
Built upon by · 0