You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In our project we are training a tabletop manipulation task involving various objects and recording the scene with a pinhole camera.
When an episode changes for a particular environment (via reset_idx), we sample locations of the objects and move them using obj.write_root_state_to_sim, and then when we capture images using the camera, we see both the objects post reset (which is what we require) as well as a lingering "shadow image" of the objects from the previous episode (see the attached images).
What is the best methodology to avoid the objects being smeared onto the next episode?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hi,
In our project we are training a tabletop manipulation task involving various objects and recording the scene with a pinhole camera.
When an episode changes for a particular environment (via
reset_idx
), we sample locations of the objects and move them usingobj.write_root_state_to_sim
, and then when we capture images using the camera, we see both the objects post reset (which is what we require) as well as a lingering "shadow image" of the objects from the previous episode (see the attached images).What is the best methodology to avoid the objects being smeared onto the next episode?
Thanks,
Tom
Beta Was this translation helpful? Give feedback.
All reactions