Closed rly closed 3 months ago
The original and labeled video files (currently represented as a list but my be changed to a set from #12) is ordered. We intend the set of cameras to align with these lists, so the nth video was generated by the nth camera. However, a set of N links (quantity: '*'
) is unordered. There is no simple and straightforward way to align these in the PoseEstimation
data type.
The proposal in #12 would change the video files to be ImageSeries
objects, which already have an optional link to a Device
that would represent the camera. That should suffice. So the devices
field should be removed from the PoseEstimation
type in the resolution of #12 and this PR will no longer be necessary.
Already implemented in a different PR.
Pending a fix for https://github.com/hdmf-dev/hdmf/issues/693