Foundation pose model input

[sarvatmak]

What input does foundation pose model takes I am confused, in document on NGC it is given
Input Types

• RGB image
• Depth image
• 2D bounding box
• CAD model
• Intrinsic matrix

Thus i am under the impression only some from the above list are enough, is it correct?
or do we have to pass all the above parameters.
Can anyone please point me towards documentation where it is specifically mentioned.

[Raffaello]

Hi @sarvatmak

Isaac ROS foundation accepts .obj models, but please help me figure out which document you are looking for.

There are some demo available:

• Tutorial to create your own 3D object mesh in .obj
• Tutorial to convert a 3D mesh from .usd to .obj using Isaac Sim
• Tutorial to simplify a 3D mesh using Meshlab

And these inputs can be from different types of sources.

Best,
Raffaello

[sarvatmak]

Actually I specially wanted to know that if I just pass RGB images (and nothing else )to the model will it be possible to detect pose?

[Ahung]

Hi @sarvatmak ,

Those inputs are required for Isaac ROS FoundationPose.

Please refer to this Isaac ROS Office Hours.

• Jul 3, 2024 - Isaac ROS FoundationPose

Best,
Ahung