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
For pose estimation. Should probably just take the keypoint of reference and subtract that from all other points. Maybe with an option to keep allocentric coordinates for the chosen keypoint so it's possible to convert back?
Could also keep the animal heading if two keypoints are given, 0 being either in-line or perpendicular to the line formed by the two keypoints.
If no reference is given, then compute the centroid and keep allocentric coordinates for that.
The text was updated successfully, but these errors were encountered:
At our movement meeting, Sofia made me realise that of course the heading of the animal in real world is also still needed, so that will have to be kept too.
I think this will need to be done in two steps:
Centering (a point is chosen, could also be based on multiple points)
Rotate everything else to maintain either a horisontal line on the x-axis or a vertical line on the y-axis
For pose estimation. Should probably just take the keypoint of reference and subtract that from all other points. Maybe with an option to keep allocentric coordinates for the chosen keypoint so it's possible to convert back?
Could also keep the animal heading if two keypoints are given, 0 being either in-line or perpendicular to the line formed by the two keypoints.
If no reference is given, then compute the centroid and keep allocentric coordinates for that.
The text was updated successfully, but these errors were encountered: