Seonglae ChoV2
Depth Anything V2
This work presents Depth Anything V2. Without pursuing fancy techniques, we aim to reveal crucial findings to pave the way towards building a powerful monocular depth estimation model. Notably,...
https://arxiv.org/abs/2406.09414
V3
Depth Anything 3: Recovering the Visual Space from Any Views
TL;DR:
Depth Anything 3 recovers the space with superior geometry and 3DGS rendering from any visual inputs.
The secret? No complex tasks! No special architecture!
just a single, plain transformer trained with a depth-ray representation.
https://depth-anything-3.github.io/
arxiv.org
https://arxiv.org/pdf/2511.10647