Credit: Apple Depth Pro
A team of engineers at Apple developed an AI-based model called Depth Pro, which can map the depth of a 2D image. The team wrote an article describing the application and its capabilities and published it on the website. arXiv preprint server. They also posted an announcement about the application on the company’s Machine Learning Research page.
Humans and other animals are able to perceive depth because the brain is able to take two images, one from each eye, and use the differences between them to determine which parts of the images are closer and which are closer. more distant. Some video cameras have done something similar to create 3D videos.
Smartphones, because they rely on a single camera for photo taking and video creation, have various hardware and software additions that help add a certain degree of depth. In this new effort, Apple engineers created a complete depth map using the original image data without relying on metadata such as camera intrinsics.
A depth map is a map created using all the pixels of an original image. Each data point on the map represents a single pixel and corresponds to the distance between the part of the image represented by the pixel and the corresponding part of the object that was imaged.
Such a card allows you to add another dimension to a flat image, giving it 3D effects. According to the team, creating a depth map can generate sharper 3D effects than those created using standard smartphone techniques.
Presentation of the network architecture. Credit: arXiv (2024). DOI: 10.48550/arxiv.2410.02073
In its announcement, the Apple team claims that applications using this model are able to produce a depth map in just 0.3 seconds when run on a computer with a standard GPU, and they can do without the types of camera data usually needed. to generate 3D effects.
By creating a model that works so quickly, Apple has opened the door to creating 3D images from a single-lens camera in real time. And that, the team notes, could have major implications for robots and other real-time mapping applications, such as those used on autonomous vehicles.
More information:
Aleksei Bochkovskii et al, Depth Pro: Clear monocular depth metric in less than a second, arXiv (2024). DOI: 10.48550/arxiv.2410.02073
Depth Pro: github.com/apple/ml-profondissement-pro
machinelearning.apple.com/research/deep-pro
arXiv
© 2024 Science X Network
Quote: Apple unveils Depth Pro, an AI application capable of mapping the depth of a 2D image (October 10, 2024) retrieved October 10, 2024 from
This document is subject to copyright. Except for fair use for private study or research purposes, no part may be reproduced without written permission. The content is provided for informational purposes only.
