Skip to content

Apple Introduces Depth Pro 2 for Sharp, Temporally Consistent Metric Depth

Apple Machine Learning ResearchFrom the source·· 1 day agoPicksDepth Pro 2: Sharp and Temporally Consistent Metric Monocular Depth
Original
Who
Apple
What
releases Depth Pro 2 models, code, and weights for metric monocular depth estimation in images and video
Quote
We release a family of models that define a new Pareto frontier in terms of metric accuracy, sharpness, temporal consistency, and runtime.

What happened

Apple researchers present Depth Pro 2, a network that estimates metric depth at native resolution from images and video. They say it combines sharp object boundaries with temporal consistency, outperforming prior work across metric accuracy, sharpness, and temporal consistency. Apple releases a family of models, along with code and weights.

Why it matters

A single network handles both photos and video while addressing the usual trade-off between sharp depth boundaries and frame-to-frame consistency; Apple also releases model weights and code for others to use.

Read next

See the full storyApple Releases Depth Pro 2 for Sharp, Temporally Consistent Metric Depth
This story is in 11/21 AI DailyRead the whole issue