 |
AI image to illustrate potential future smart glasses (Image credit: Ideogram 2/Future AI) |
Every week brings something new in AI development that advances the field; this week's innovation originates from a small Cupertino tech startup.
While all eyes are on Apple Intelligence and its imminent release which will bring context-specific AI capabilities to everyday usage, the firm has also shown off a new AI model dubbed Depth Pro.
As the name indicates, this new artificial intelligence model will map the depth of a picture in real time. What's more fascinating is that it can accomplish this with regular home computer hardware; an Nvidia H100 is not needed.
Although Depth Pro is a research model and not necessarily something Apple is putting into production, it would undoubtedly help the firm enhance the usefulness of augmented reality or even the Vision Pro's if we ever get a pair of Apple Glasses.
 |
(Image credit: Apple Depth Pro) |
Apple released an incredible ML Depth Model yesterday that creates a depth map in *meters* from a single image
— Alex Krause (a19grey.eth) (@a19grey) October 6, 2024
I built a demo to play with it
- Added ability to download depth map in meters
- AND can generate a real-scale 3D object file of the scene
(forked from a space by… pic.twitter.com/XdbtqN9Dp4
Apple's new methodology creates "metric depth" by estimating both absolute and relative depth. The image and its data may subsequently be utilized in a variety of ways.
When a user takes a photo, Depth Pro creates precise measurements between elements in the image. Moreover, Apple's model ought to steer clear of contradictions such as misinterpreting the foreground and background of a picture or believing the sky to be a part of the backdrop.
Uses for Apple's Depth Pro AI: Transforming 3D Space Capture
The possibilities, Terminator 2 apart, is practically unlimited. Drones, robot vacuums, and autonomous automobiles (which, ironically, resemble Apple's shelved product) may all utilize precise depth sensing to enhance object avoidance, while online furniture retailers and augmented reality technology might assist arrange objects in real or virtual spaces more correctly.
Depth perception might also help medical technology by facilitating better interior organ mapping and anatomical structure restoration.
It might go full circle, too, more precisely helping convert photos to video using generative AI like Luma Dream Machine. In order to help the video model better grasp how to manage object placement and motion in that area, the depth data would be sent to it together with the picture.