International
The Morning Science

What science found today

‹
 

Research

Depth Pro 2: Sharp and Temporally Consistent Metric Monocular Depth

Apple ML Research ·

The Brief

  • Apple presents Depth Pro 2, an approach for estimating dense metric depth from images and video at full resolution.
  • The authors say prior monocular depth methods commonly trade off temporal consistency against boundary sharpness.
  • Depth Pro 2 is presented as sharp and temporally consistent metric monocular depth estimation.

Why it matters

It targets a known trade-off in monocular depth estimation, so video-based depth estimates could be both sharper at object boundaries and more stable across frames.

Who’s affected

  • computer vision researchers
  • robotics and augmented reality developers working with monocular depth
  • Depth Pro 2
  • #monocular depth
  • #depth estimation
  • #computer vision
  • #video
  • #apple research

We present Depth Pro 2, an approach to estimating dense metric depth for images and video at full resolution. By analyzing prior methods for monocular depth estimation, we find that they commonly trade off temporal consistency and boundary sharpness: multi-frame methods favor…

Timeline

  1. Apple Machine Learning Research publishes Depth Pro 2 paper

  2. Apple Machine Learning Research

    Depth Pro 2: Sharp and Temporally Consistent Metric Monocular Depth

How do you feel about this?

0 Comments