Apple releases Depth Pro 2 for improved monocular depth estimation
In short: Apple has introduced Depth Pro 2, a new approach to monocular depth estimation that processes images and video at native resolution. The model resolves the traditional tension between sharp object edges and frame-to-frame temporal stability. It provides dense metric depth in meters without requiring camera metadata.
This summary was generated automatically by AI from Apple Machine Learning's publication. It is our own text, not a copy of the original — facts, figures and quotes belong to the source, linked above and below.
What changed?
- 1Unified architecture for both single images and video
- 2Native resolution depth map generation
- 3Simultaneous improvement of metric accuracy, edge sharpness, and temporal consistency
Why it matters
For computer vision pipelines, this model offers a way to perform high-fidelity depth mapping on standard camera inputs. By providing temporal consistency for video feeds, it simplifies tasks requiring spatial awareness in dynamic environments without needing complex multi-frame pre-processing.
Sources
- Apple Machine LearningOfficialPrimary sourceOriginal article →„Depth Pro 2: Sharp and Temporally Consistent Metric Monocular Depth“20 Nov 2026, 02:00
- Published by source
- 20 Nov 2026, 02:00
- Found by our system
- 20 Nov 2026, 21:38
- Summary generated
- 21 Nov 2026, 21:09
This article was written by AI from the original source. Facts, numbers and prices come from the source; missing values are marked “Not specified”. Legal notice, copyright and privacy