Noise Generation

Hi everyone, first time posting.
I’m using omniverse to generate a dataset to train a model for image segmentation using both image and point cloud.
Training goes excellent, but I’m having issues with the sim2real part.
I think it’s probably because both the camera and the point cloud generated by the stereo (I’m using a real sense stereo camera) have intern noise.
I can easily generate noise for the image, but is there an easy way to add noise to the point cloud inside Omniverse?
Also, follow-up question, I’ve seen that the ray generation for the point clouds is physics-based. For the same project, I was wondering if there was a way to add particle noise (dust or smoke) into the mix or to change the ray generation into a render-based one.

Thank you in advance for reading

Hello! I’ve reached out to the dev team for some assistance. Thanks for reaching out to us!

Update from the developers:

Pointclouds are store inside usd as usdGeomPoints, you could add random values to the elements as you like - how to add the noise really depends on your application.

Thank you very much, will try that.