Hello,
Thanks for your work. I have a question regarding data collection that you uploaded in the form of recordings in the google drive. How did you generate the hdf5 during real data collection ? Was it the position and orientation stream that you captured from every joint (hdf5 file) along with an svo (left and right pair)? If so, how did you ensure that the joint positions are pixel aligned ? I mean if they are directly coming from controllers from IK streaming, there should be a lag between the joint pose in pixel space vs the joint pose showed by the controllers. Which one do you record for hdf5 file and at what hz (zed streams at 30 hz I suppose, so svos should be at the same rate I believe)?
Thanks.
Hello,
Thanks for your work. I have a question regarding data collection that you uploaded in the form of
recordingsin the google drive. How did you generate the hdf5 during real data collection ? Was it the position and orientation stream that you captured from every joint (hdf5 file) along with an svo (left and right pair)? If so, how did you ensure that the joint positions are pixel aligned ? I mean if they are directly coming from controllers from IK streaming, there should be a lag between the joint pose in pixel space vs the joint pose showed by the controllers. Which one do you record for hdf5 file and at what hz (zed streams at 30 hz I suppose, so svos should be at the same rate I believe)?Thanks.