Skip to content
RobotWorld
← Back to Videos

DeFM: A Depth Foundation Model for Robotics

Loading video
Loading video

ETH Zurich's DeFM is a self-supervised depth backbone trained on 60M images, released as 12 models for navigation, manipulation and locomotion.

Depth perceptionFoundation ModelsSelf-Supervised LearningRoboticsETH Zurich
Category: perception
Author: @robotsdigest
Date: 2026-10-28T00:00:00
Duration: 136.802s
Source: https://x.com/robotsdigest/status/2115521436513395118
Reference: https://arxiv.org/abs/2601.18923

Related Videos

SSIO: Self-Supervised Inertial Odometry Without Labels
0:49

SSIO: Self-Supervised Inertial Odometry Without Labels

@QiuYuhengQiu · 24 views · 2026-10-28
SSIOInertial Odometry
ROS 2 3D LiDAR Localization on Prior Maps

ROS 2 3D LiDAR Localization on Prior Maps

@rsasaki0109 · 145 views · 2026-10-16
ROS2LiDAR Localization
microSLAM: Monocular SLAM for Interactive 3D Environments
3:30

microSLAM: Monocular SLAM for Interactive 3D Environments

@ZenoInMotion · 771 views · 2026-09-10
SLAMMonocular Vision
LingBot-Depth 2.0: Transparent Object Perception
0:30

LingBot-Depth 2.0: Transparent Object Perception

@robbyant_brain · 482 views · 2026-07-31
reflective objectPerception
Early access to NVIDIA Isaac GR00T N1.7 is here 🎉 — an open, commerci...
0:10

Early access to NVIDIA Isaac GR00T N1.7 is here 🎉 — an open, commerci...

@NVIDIARobotics · 572 views · 2026-04-17
NVIDIA GR00THumanoid
Astribot Lumo-2 — 20+ household tasks, one foundation model, real robot
4:57

Astribot Lumo-2 — 20+ household tasks, one foundation model, real robot

@XRoboHub · 545 views · 2026-07-16
VLAAstribot

As an Amazon Associate, we earn from qualifying purchases.