Share via

About conversion processing in AzureObjectAnchors

IEABJKL 1 Reputation point
2022-11-03T08:49:25.93+00:00

I have a question about Azure Object Anchors.

In AOA, I need to send my own model to the cloud for conversion.
It is ambiguous to say that AI is used to extract feature points and shapes.
Specifically, what kind of processing is being performed during conversion?
It is ambiguous to say that AI is used to extract feature points and shapes.

I apologize if this has already been explained.

Azure Object Anchors
Azure Object Anchors
An Azure mixed reality service that automatically aligns and anchors three-dimensional content to objects in the physical world.
17 questions

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. Nathan Manis - MSFT 481 Reputation points Microsoft Employee
    2022-11-15T18:16:37.433+00:00

    Hi @IEABJKL ,

    Thank you for the post. For the model, Azure Object Anchors will do the conversion in the cloud to use for mapping. We do not detail the extraction algorithm. We do state in the FAQ that geometry is used in the algorithm.

    We do note that the model that is being uploaded should be within these parameters for the conversion process. This helps provide some details that can be inferred that are relevant to the algorithm like size, color, etc:

    https://zcusa.951200.xyz/en-us/azure/object-anchors/faq#product-faq

    We recommend the following properties for objects:

    • 1 -10 meters for each dimension
    • Non-symmetric, with sufficient variations in geometry
    • Low reflectivity (matte surfaces) with bright color
    • Stationary objects
    • No or small amounts of articulation
    • Clear backgrounds with no or minimal clutter
    • Scanned object should have 1:1 match with the model you trained with

    Thanks,
    Nathan Manis

    1 person found this answer helpful.
    0 comments

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.